Contents

Introduction
Characteristics of python programming
History of Python language
What is Python?
The advantages of Python
The disadvantages of Python language
The design limitations
Common Terms You Should Know with Python
Python Language Structure
Getting Started with Python
Learning the Basics of Python Programming
Flow of Control
Comments
Escape sequences
Variables and What They Do in Python
Understanding the scope of a variable
Modifying values
Python programming 2
How does Python Execute a Program
Survey of a Simple Program
Some of the Basic Commands, Variables, Statements, and Other Things that you can do
Allocating values to your variables
Multiple assignments
Keywords
Operands and operators
Understanding the decision control structure
Loop Control Statements
Functions
Lifetime and scope variables
List comprehension
What is machine learning?
Sorts of Systems of Machine Learning
What is Supervised Machine Learning?
Classification:
Unsupervised Learning
Association
Semi-Supervised Machine Learning
System Testing
System classification
Line Plot
Images
How to train a model
Gradient descent
Batch gradient descent
Stochastic gradient descent
Mini-batch gradient descent
Regularized Linear models
Mass function of a binomial distribution:

Introduction
Most computer language seems terrible to programmers that are yet to
start coding, you might have taken a look at a couple of the popular
programming languages, for example, C++ or Java and been somewhat
humiliated by what you saw. The pages may have been filled to the edge with
letters and images that you simply didn't comprehend, and you got
disappointed and just needed to leave. Numerous individuals are scared of
programming and feel like it is just unreasonably hard for them. In any case,
with the Python programming language which is known as home for all, as it
is recommended by programmers around the globe as a good language for
beginners to learn, you will find out that it doesn’t take much for one to read
and understand programming. This book is going to give you the
fundamentals that you need to acquire for a smooth Python programming
journey.
The world of computer has acquired various kinds of people. Some are
keen on bringing in cash by creating their own programs to offer to other
people. Why few do play and learn various things about how the system will
work. Also, many have dedicated their lives to programming, making it the
item that brings home their salary every month whether they work at fixing
systems, work in a company to protect the computer, or doing some other
types of computer technology. With regards to computer technology, nothing
will be straightforward. Previously you can even get a program to chip away
at the computer; it needs to get the right code to make it work. There are a
few choices for program creation that a programmer may choose including
Java, C++ and Python. Here we will investigate some things about Python
and why it is regularly favored over the other programming alternatives.
Before you can begin using Python to assume control over your programming
needs, it is essential to begin studying it and the whole incredible advantages
you will get when utilizing this program. Python is an elevated level
programming device, which implies that it is always simple to utilize and
peruse, even as a learner. The way of thinking behind the code is lucidness
and it has a kind of sentence structure that permits the developer to
communicate their ideas without having pages of code alongside it.
Readability and simplicity are python’s top priorities, which is an
extraordinary language for amateurs to begin on the grounds that they will
really have the option to read and comprehend the code they are placing in.
With different alternatives to programming, they may need to invest a great
deal of effort trying to get the code just right, including numerous different
objects to get it to work. Be that as it may, with Python, it is kept simple and
you will find that it is easy to understand what you are doing.

Characteristics of python programming

In computer world there are many coding programs that are there for you to
use but amongst them python stands tall, this is because python as a coding
program has some unique features that distinguished it from other programs.
Below are the characteristics of python

1. Well designed syntax: python has a rich grammar which will make
the projects so natural to peruse, the language is simply English
which makes it work without a great deal of bugs. Python has as a
huge library that will work with other programming tasks, for
example, evolving documents, scanning for text, and connecting
with web servers.
2. Python is extremely intuitive. This makes it simpler for you to try
out little bits of code to check whether they work. You can
likewise organize it with a development environment called IDLE.
On the off chance that you might want to extend the programming
language, it is extend into different modules like C or C++.
3. Availability: Python programming can be run on any unit
including Unix, Linux, Windows, and Mac OS X.
4. Programming for all: Python is free. You won't need to pay
anything to download and install Python in any device. You can
likewise make alterations and redistribute this software. It is under
a license, however it is an open source license so others can utilize
it as well.
5. Pro features: Despite the fact that Python is a straightforward
programming language, it contains some advanced features such as
list comprehensions and generators. The software is sensitive to
mistakes (errors are easily detected), since instructions are
progressively composed, when you combine types that don't
coordinate, it will raise an exception for you to take note. You can
arrange the codes into bundles and modules if necessary. There is a
wide assortment of fundamental data types that you can browse
like dictionary, strings, lists and numbers
History of Python language
 The Python language was created in 1990 by Guido Von Rossum at
Stichting Mathematisch Centrium in the Netherlands. The language itself has
been actually developed by a large team of volunteers and is open sourced.
At that point, van Rossum was dealing on a project with the Dutch CWI
research foundation, that was later closed. Van Rossum used some basics of
this new computer language (ABC language) had the option to utilize a
portion of the basics of this new language, known as the ABC language, so as
to deal with Python.
In early 2000, Guido and some of the key python development team
went beopen.come to form the BeOpen PythonLabs team. In the same year,
the PythonLabs team moved to Digital Creations which is now known as
Zope Corporation.
Python 2.0 was released on the 16th of October in the year 2000 with features
including a garbage collector which helps maintain memory handling related
issues in programming. The great thing about Python was that it is backed by
a community and it has a transparency behind the users that utilize the Python
language.
Python 3.0 which was a major backwards-incompatible version was released
on December 3rd 2008 after a long time of testing. The major features of
Python 3.0 also have been backported to backwards-compatible Python 2.6
and 2.7.
Python Version Release Dates:
Python 1.0.0 (series) - January 1994- September 5, 2000
Python 2.0.0 (series) - October 16, 2000- July 3, 2010
Python 3.0.0 (series) - December 3, 2008-till date
The fundamental quality of this language is that it is extremely simple to
reach out upon to make progressively complex, or keep basic, and it has the
option to help various levels of intelligence. Both of these were significant
during the days when individual computers were getting into the mainstream.
Furthermore, since Python was made to interact with various file formats and
libraries, it turned into a hit also. Python has developed a considerable
amount since its commencement and more tools have been added to make the
programming all the more working. Notwithstanding making Python simple
to utilize, van Rossum team has been working at various stages to see that
python meets everybody’s need. Using Python language to code can make
things simpler and assists with disposing of a portion of the feelings of
unease related with the unpredictable computer codes since it doesn't look so
alarming. Throughout the years, van Rossum chose to make Python open to
all. This permitted all to get entrance and make changes to Python so that if
something happened to van Rossum, all would not be lost. On account of
having Python open sourced, Python 2.0 was released during 2000 to make it
community based and to have a straightforward improvement process. There
are a couple more up to date forms of Python 2.0 despite everything being
utilized, however Python 3 the latest release stormed the world. Python 3 was
released in 2008. It isn't just an update to the program, however a total
change. While there are a ton of incredible features that accompany this
version, non is compatible with other so you will need to settle on a decision
between Python 2 and Python 3. To make things simpler, the software
engineers made a little marker inside the program that would demonstrate a
programmer what should have been changed between the two programs when
uploading. In spite of this, most have stayed with Python 2.0 for the time
being.

What is Python?
Python is generally an object-oriented, high-level programming language
with dynamic syntax. Python has been distinguished because of the increased
efficiency that it provides since the edit-test-debug cycle is incredibly fast
compared to other programming languages.
The Python language has a simple, easy to learn syntax which is empowered
with many English words for easier readability and assists for increased
productivity and efficiency. When programming in Python, it feels more like
you are writing out the solution to a problem in your own thoughts rather
than trying to refer to ambiguous symbols that are needed in the language to
give to certain functionalities. Python language can be used to automate
measurements and process data interactively. Python is able to handle large
databases and compute large numbers with much ease compared to many
other programming languages. It can be used as an internal scripting
language so that it is executed for certain functions by other programs. By
learning Python, you will be able to write complex software like intricate web
crawlers. It is truly an all-purpose language. The great thing about Python is
that it has a giant library for web crawling and is used a lot for its capability
in scraping the web. A web crawler is simply a program that can navigate the
web depending on the parameters set out for it and scrapes content that you
would like for it to scrape. All in all, Python can be easy to pick up whether
you’re a beginner in programming or an expert on other languages. It’s a fast,
friendly and easy to learn language but don’t mistake that for its powerful
nature.

The advantages of Python

Python is presumably a standout amongst other programming languages that
you can pick to utilize. Beginners are going to cherish that it is so natural to
turn on to this program and begin composing their own codes, even without
much experience, and there is bounty to appreciate when you are an expert,
or a specialist, too. Below are some of the benefits one can derive from using
python.

1. Simple to use and peruse

With regards to programming language, there are none that are as simple to
utilize as Python. Different languages are somewhat cumbersome and
difficult to study; you will notice that they have huge amounts of bracket and
even words that you can’t understand which is sufficient to discourage
somebody who isn't used to programming at all since all the words look
somewhat scary. Python is somewhat unique. Rather than all the insane
brackets, it uses indentations, causing a simpler to peruse page that isn't such
a wreck. Rather than words that you can't comprehend, it utilizes English.
The other extraordinary characters are kept at minimum with the goal that
you can easily study a page of code without any stress. This is one of the
simplest programming tools that you can use excel or even to find out more
about programming.
 2. Effectively present on certain PCs
At times, Python is as of now present on your PC. Macintosh OS X systems
just as those with Ubuntu will as of now have Python preloaded. You will
just need to download a text interpreter to begin. As far as utilizing Python on
Windows PCs, you should simply download the program. Python works with
these PCs, regardless of whether it isn't installed from unset.

3. Uses English as its core language

Since English is the language that this program is built on, it is extremely
simple to peruse. There aren't a ton of words that you won't get and you won't
need to spend the whole day trying to decode the commands. The program is
all in English and you will adore how straightforward this can make things.

4. Can work with other programming languages

Before all else, you will probably just utilize Python all alone. It is an
extraordinary program to learn with and develop with. In any case, after some
time, you may conclude that you need to have a go at something new that
Python can't do all alone. Fortunately, Python is ready to work with a few
other programming language, for example, C++ and JavaScript, so you can
play around, get familiar with some more, and truly get the code that you are
searching for, regardless of whether Python can't accomplish all the work.

5. Can try out things with the interpreter

At the point when you download Python, you will need to download a text
interpreter as well. This will make it simpler for Python to peruse your data.
You can utilize common products that are already on your PC, for example,
Notepad from Windows or search for another interpreter that might be
somewhat straightforward. When you select the interpreter that you might
want to utilize, the time has come to get to work composing the code. A
portion of the individuals who are new to coding may feel stressed about
attempting to get the code to work. This is another spot where Python can
make things simpler. It will have the option to take the words that you are
composing and spit them back, with the assistance of the interpreter, in only a
couple of moments. You can test what you are doing while you are chipping
away at it! There are such a large number of advantages of utilizing the
Python program. beginners are going to love the delightful way promptly
accessible this program is and that it is so natural to become familiar with
some of the basic commands right away. Indeed, even the individuals who
have been programming for some time will be intrigued by how this all
functions!

The disadvantages of Python language

While there are a great deal of motivations to adore Python, it is critical to
understand that there are a couple of disadvantages in using python that you
should look out for. These drawbacks include:

1. It doesn't have a great deal of speed

For the individuals who are hoping to work with a program that has a great
deal of speed, Python may not generally be the best choice for you. It is a
deciphered language so this will slow it down when compared to different
alternatives that are aggregated languages. In any case, it relies upon what
you are translating. There are sure benchmarks with the Python code that can
run quicker utilizing PyPy contrasted with different codes. Fortunately this
issue with a moderate speed and Python is being cured. Software engineers
are trying their best to develop a faster interpreting speed.

2. Not accessible on most mobile browsers

Python is an incredible alternative to utilize in the event that you have a PC.
It is accessible on numerous desktop and server platforms to assist you with
making the code that you are searching for. Be that as it may, it isn't all set
into mobile devices. Since there is such a major increment in income and
individuals going into the versatile business, it is tragic that this programming
language hasn't stayed aware of the patterns like others. Maybe later on
Python will choose to go into the future and build up a software that will have
the option to function admirably with different cell phones. Up to that point,
software engineers should be happy with utilizing it on their PC and
desktops.

The design limitations

On the off chance that you are hoping to work with a program that has a ton
of designs the Python program may not be the correct alternative for you. The
design language isn't up to what you will discover with other alternative
languages. Since you are working with a program that is progressively
composed, it takes all the more testing and can have more mistakes that will
possibly show up when you are running the program. The worldwide
interpreter lock implies that you can just have one thread, access the internals
of Python at once. This may not be as significant any longer since it is simple
to bring forth the errands out to various procedures, however the plan isn't as
decent as other alternatives that you might want. A decent method to work
with the task is to recall that space is significant with Python. Other
programming languages are going to utilize a ton of bracket to show the
distinction in lines and data inside the program, yet Python will depend on
spaces. Try to be cautious with utilizing this to maintain a strategic distance
from issues and mistakes that can come up.
Python can be probably the best program that you use to compose your own
codes furthermore, have a fabulous time. While there are a great deal of
advantages to utilizing this program, particularly contrasted with a portion of
different ones that aren't as simple to peruse, it is essential to comprehend
both the positives and negatives of every alternative previously you hop in!
Common Terms You Should Know with Python
Before you get excessively far into your programming with Python, it is
essential to note a portion of the words that can make the programming
simpler to comprehend. This part is going to set aside some effort to take a
look at the extraordinary words that are regular in Python programming, and
which we do discuss a bit in this book, to help stay away from some disarray
and to assist you with beginning with your first code.
Class: this is a format that was utilized for making user defined objects.
Docstring: this is a string that will show up lexically first expression inside a
module, function, or class definition. The item will be accessible to
documentation tools.
Function: this is a block of code that is conjured when utilizing a calling
program. It is best utilized so as to give a calculation or an independent
service.
IDE: this represents Integrated Development Environment for Python. This is
the essential interpreter and editor environment that you can utilize alongside
Python. It is useful for the individuals who are simply starting up their
programming life and can work for those on a careful spending plan. It is just
simply a unique code and doesn’t accommodate much space
Immutable: this is an object inside the code that is relegated a fixed value.
This could incorporate tuples, strings, and numbers. You can't change the
object and you should make another object with an alternate value what's
more, store it first. This can be useful at times, for example, the keys in a
dictionary.
Interactive: one good thing about Python is that it is so interactive. You can
experiment with the interpreter and see how it responds to the new ideals,
which is a short way to develop your programming skills.
List: this is a datatype inside Python that is implicit. It contains a mutable
sequence of values that are arranged. It can incorporate inmutable values of
numbers and strings too.
Mutable: these are the objects that will have the option to change their value
inside the program, however which can keep their unique id().
Object: inside the program, this is any information with a state, for example,
a value or a quality, just as a characterized conduct, or a strategy.
String—this is one of the most essential sorts that you will discover in Python
that will store the content. In Python 2, the strings will store text so that the
string type would then be able to be utilized to clutch binary data.

Triple quoted string: this is a string that has three occasions of either the
single statement or the double statement. It could have something like '''I love
tacos'''. They are utilized for some reasons. They can assist you with having
double and single statements in a string and they make it simpler to go over a
not many lines of code without issues.
Tuple: this is a datatype that has been incorporated with Python. This
datatype is an immutable arranged sequence of values. The sequence is the
main part that is immutable. It can contain some variable qualities, for
example, having a dictionary inside it, where the value’s can change.
Type: this is a classification of data that is represented in the programming
languages. These sorts will vary in their properties, they have mutable and
immutable option, just as in their capacities and techniques. Python
incorporates a couple of these including dictionary, types, tuple, list, long,
integer floating point and string.
In order to install Python on system, you have to download the following:
1 Python IDE
2. Python Interpreter.
The download of these two tools will facilitate the process of becoming a
Python programmer. An IDE is a packaged application program used by
programmers because it contains necessary tools in order to process and
execute code. An IDE contains a code editor, a compiler, a debugger, and a
graphical user interface (GUI). There are different types of IDE but the most
popular among them is PyCharm for Python. For you to use the PyCharm
IDE from JetBrains, you must first install the Python Interpreter (IDLE
version 3.4) on your system so PyCharm is able to detect Python on the
computer. For you to download the python interpreter use the link bellow:
https://www.python.org/downloads/ once you have open the link you will
find download then, click download and a file will be downloaded
accordingly. Once you have this file, execute it and follow the instructions on
the wizard to install the Python Interpreter. After the installation of python
interpreter you can proceed to downloading the IDE PyCharm. In order to
install PyCharm, visit the link below:
https://www.jetbrains.com/pycharm/download/ and click the “Download
Community” button and continue to download PyCharm. During this process,
it will automatically detect the Python 3.4 IDLE but in rare cases you might
need to specify the directory that you installed Python in.
Once you have completed the installation of PyCharm and the Python 3.4
IDLE, go to File New Project. You will reach this screen:
You can rename the “untitled” portion of the Location field to a specific
project name. Then click the “Create” button and you will end up being in an
empty project. The next thing we must do is create a new Python file which
you can do by going to: File New Python File Set the name of your python
file accordingly and it will show under your project name in the Project
Explorer. Now double click on the file and you will be met with a page where
you can start coding Python.

Python Language Structure

We will learn how Python as a programming language is structured through
the following sample code: import math
class Program:
def Execute(self):
x = math.sqrt(25) print(x)
run = Program()
run.Execute()
For you to execute the code, you must use the shortcut alt + shift + f10 or
execute it using the Run menu item in the main navigation bar in PyCharm.
The first line import math is using a unique keyword import which allows the
programmer to import tools from the Python library which aren’t built-in by
default when coding. After the keyword import - the programmer specifies a
specific directory that is within the Python library. In this case, we specified
the math directory, if we wanted to specifically specify something within the
System directory; we would add a dot separator (period/full stop) and then
added a sub directory name. If the library’s folders are more complex, you
must use notation for instance: import bs4 from BeautifulSoup. The
BeautifulSoup library is an example of where we want to import the bs4
directory, which is why the “from” keyword is used in the beginning of the
program. In our case, we imported the math module because it contains
functions. Many libraries contain functions that are
able to be called so we don’t have to code our own functions from scratch. In
this case, we used the square root function from the math module. In this
case, we used math.sqrt() and then printed out the value it returned relative to
the input it received.
We first typed in math to indicate the module we are using and then a dot
separator to reference the function within the module which is the sqrt
function. The sqrt function takes in a value in its parameter or parenthesis.
While the sqrt function computes the number 25 and returns the value of 5,
this value is stored in the x variable which is why it has an equal sign to show
that it is equal to the value that is computed from the sqrt function.
Now as we run through each statement, bear in mind that in order to run the
next line, we must have a separate line for each statement. The compiler
(process to convert language in to readable code that the computer can
understand like binary) will then know when statements end and when they
start by the use of line breaks.
class - A class can be thought of as a “section” of program. For example:
Section: Everything in here is the contents of the section. Again, you will
gain a better understanding how classes are useful in Inheritance and
Polymorphism.
Program - This second element of this important line is “Program” which
simply is a custom name that the user can redefine. You can call this
anything, as all it is doing is giving the “Section” or “Class” a name. You can
think of it this way: Section Name:
Another thing you must be afraid to look at is the colon: “:” and “}” - what
the colon does is simply tells the compiler when a specific section starts. This
could be thought of as when someone is writing an essay and they have to
start their sentences with the appropriate words or indent their paragraphs.
One thing to note is that the spacing in code does matter. If you are creating
aclass, the content of that class will be indented once using the tab key in
order for it to interpret that the code that the class has authority over is the
tabbed code underneath it. Any code that is not tabbed underneath class is not
part of that class. This goes the same for any function you create as well. As
shown in the above example, a function is defined with a colon and tabbed
content underneath it to indicate that it is part of that function. The way
Python works is that it has classes, which are sections and functions which
are subsections of the class. These classes can be declared as they are
declared in the main program after the class is declared and then called with
the functions that they carry within them. In this case, it Execute.

Getting Started with Python

Since we know a portion of the advantages of picking this program, the
time has come to begin with it. Before you can become familiar with some of
the extraordinary procedures that are expected to cause this program to make
code for you, the time has come to set up the environment. For the
individuals who have a computer with Mac OS X or Ubuntu, you will as of
now have Python installed in the system this can make things simpler to
begin as you will simply need to click on the symbol to begin. Windows PCs
should install Python. While Python works fine and dandy on Windows PCs,
it doesn't come preinstalled so you should do this.
These steps works for all Windows version:
Download Python—you can pick between Python 2/ Python 3. Both are
awesome alternatives; it just relies upon which one will take care of your
business for you. Click to run the Python Installer. At the point when you get
to the options, decide to customize Installation. You will see a box pop up.
Click on each box that is under Optional features and afterward proceed. On
the following screen, search for the Advanced Options and afterward pick
where you might want to have Python installed. When you have gotten this
far, the following part is to set up your PATH variable. This will permit the
user to include directories for all the segments and bundles that are required.
To do this progression:

Open up the Control Panel on your system.

Look up Environment.
Under System environment variable, click on Edit. At that point
click on environment Variables.
You may need to search a bit for the following part, however search for User
Variables. You can then either make another one or alter an existing path. To
make another path, select PATH as the name and add it to the directories that
are there. Ensure that every Variable Values is separated with a semicolon. In
the event that you need to alter your existing path, you have to ensure that
each value is on an alternate line. Click in the event that you need to alter
your existing path, you have to ensure that each value is on an alternate line.
Click on New and afterward put your directories on various lines. Presently
you can open your command prompt. To do this click on Start at that point
Windows System and afterward Command Prompt. At the point when the
command prompt opens, you can type in "Python." This will load up the
Python interpreter. You would then be able to type in Exit and hit Enter to
return to the command prompt.

Text editor
You won't have the option to program Python without having the text editor
on your Computer. On the off chance that you are utilizing Windows, the
Notepad capacity will work. Ensure that you are not utilizing Word because it
isn't viewed as an editorial manager furthermore, your code won't save on the
system appropriately. In the event that you are considering getting a variant
of Notepad, you will see that Notepad ++ is the best one to use on a Windows
and Text Wrangler is the best to use for Mac. To set them up, fellow the steps
bellow:

Windows

Download and afterward install Notepad ++

When it is downloaded, open up the settings and click on
Language Menu and Tab Settings
Check the box that is close to Expand Tabs
Ensure that the value is at 4.
Click again to Close.
Macintosh

Download and afterward install Text Wrangler

You won't have to register to install the product, simply click
Cancel if there is a box that pops up requesting this.
You can as well adhere to the instructions that appears on the
screen to set this editor.
When the program is on your PC the time has come to become
familiar with the coding and, functions that you can appreciate on
Python.

Getting IDLE
While you are setting up Python, ensure that you download the integrated
Development and Learning Environment. This ought to download alongside
Python on the off chance that you are setting it up, however ensure to look
into this while you are experiencing the procedure. This is the environment
that you are going to work with when you are on Python and it can make
things simple. On the off chance that you try not to need to play with finding
another environment or you need to make the process as simple as workable
for you as a novice, this is the alternative for you.
The primary features of using integrated development learning environment
with your Python programming include:

integrated debugger with continual breakpoints, call stack

visibility, and stepping to make things simpler
Python shell that will highlight the syntax
Multi-window text editor that can help with the indentation,
highlighting and, finishing the code.
Presently, you can decide to utilize another environment, similar to those that
we examined above if necessary, however since this one frequently comes as
an alternative with Python and it is intended to function admirably with this
system, there are numerous individuals who decide to go with this choice.
That being stated, there have been a few issues before with IDLE
experiencing difficulty focusing, won't copy a few things, and some clients
try not to like the interface design. You might need to evaluate this program
ahead of time and check whether it is the correct one for you or on the off
chance that you might want to utilize one of the alternatives above.
Getting Python set up on your PC is a really simple procedure. There are
effectively a few sorts of PCs that have the programming language well
present so you won't need to do any work and the rest of them basically need
a speedy download to finish. You can stand by only a short measure of time
to get Python on your PC and afterward you are a done and can now try out
some codes.

Learning the Basics of Python Programming

Presently the time has come to become acquainted with more about Python
programming and how you can make it work for you. You should gain
proficiency with more about the various watchwords and the factors that
accompany Python so you can compose the code that you need and cause the
program to act with a specific goal in mind. Now let’s look into some of
these essentials of Python programming so you can start creating your new
code immediately.
Keywords
At the point when you are chipping away at another PC coding program, you
are going to notice that every scripting language will have certain keywords.
These are the words that are intended for a particular command in the
language and you should ensure not to abuse them elsewhere, in the event
that you do (utilize these words in different pieces of your code), you may
wind up with an error alert or the program not working appropriately. The
keywords that are saved for Python include:

And
Pass
Or not
Nonlocal
None
Lambda
Is
In import
If
Global
From
For
Finally
False
Except
Else
Elif
Del
Def
Continue
Class
Break
Assert
As
 Yield
With
While
Try
True
Return
Raise
Identifier Names
At the point when you are making another program in Python, you are going
to work on making many entities, a mix of functions, classes, and variables.
All of these will be given a name that is otherwise called an identifier. There
are a scarcely any standards that you have to follow while framing an
identifier in Python including:

It ought to contain letters, either capitalized or lowercase or a mix

of the two, numbers, and the underscore. You ought not to have
any spaces inside.
The identifier can't begin with a number
The identifier can't be a keyword and it should exclude one of the
keywords inside.
In the event that you defy one of these norms, the program will close on you
and will show a Syntax error. Likewise, you have to take a shot at making
identifiers that are intelligible to the natural eye. While the identifier may
sound good to the PC and get through without causing issues on the PC, a
human is the one who will peruse the code to utilize it themselves. On the off
chance that the natural eye doesn't comprehend what you are writing in a
specific spot, you could run into certain issues. A few of the standards that
you ought to follow while making an identifier that is destined to be
intelligible to the natural eye include:

The identifier ought to be spellbinding—you should select name

that is going to describe what is inside the variable or will depict
what it does.
You ought to be cautious while using abbreviations that aren't
fundamental since these consistently make things that are difficult.
While there are a great deal of ways that you can work out your code, you
ought to be cautious and stick with one guideline all through. For instance,
both MyBestFriend also, mybestfriend work in the coding scene, yet pick one
that you like and do it the same each time that you work in the program to
keep away from disarray. You can additionally include underscores into this
or numbers, simply be cautious that you keep things constant.

Flow of Control
When dealing on the Python program, you will work out the statements in a
rundown design, much the same as you would when working out a shopping
list. The PC will begin with the principal instruction before working through
each of them in the order that you make them appear on the list. So you
should work out the controls that you need simply like you would for your
basic food item shopping list to ensure that the PC is perusing it
appropriately. The PC will just quit perusing this rundown once it has done
the last instruction to finish. This is known as the flow of control. This is a
significant method to begin. You need to ensure that your flow of control is
even and smooth for the computer to peruse. This will make it simpler to get
the program to do what you want without much difficulty and do ensure that
the computer program is free from any issue.
Semi-colons and Indentation
If you examine some other computer languages you will notice that there are
lots of brackets used to organize the various blocks of code most times used
to start and end statements. This encourages you to make sure to indent the
code blocks in these languages to make the code simpler to read, despite the
fact that the system can peruse the various codes without any indentation.
This sort of coding can make it extremely hard for one peruse. You will see a
great deal of superfluous information that is required for the PC to peruse the
code, yet can make it hard on the natural eye to understand this. Python
utilizes an alternate method of doing this, generally to help make it simpler
on the natural eye to peruse what you have. You are going to need to indent
the code for this to work. A case of this is:
# this function definition starts another block
def add_numbers (a, b):
c= a + b
# as is this one
return c
# this function definition starts a new block
if it is Saturday
print (It's Tuesday!"
# and this one is outside the block
print ("print this no matter what.")
Moreover, there are a ton of languages that will utilize a semicolon to tell
when an instruction end. With Python however, you will utilize line ends to
tell the system when an instruction will end. You will have the option to
utilize a semi-colon if you have a couple of instruction that are on a similar
line, yet this is regularly thought about inappropriate behavior within the
language.
Letter Case

Most system languages will treat up ercase and lowercase letters equally, yet,
Python is one of the main ones that will be case sensitive. This implies that
the lower case and upper case letters will be dealt with differently in the
system. Remember too that all the reserved words will utilize lower case with
the exception of None, False, and valid.

These fundamentals are going to make it simpler to begin on the Python

programming. You have to set aside a touch of effort to experience the
program all together to get acquainted with it. You won't have to turn into a
specialist, yet getting acquainted with some of the text interpreter and few
other pieces of the program can make it simpler to utilize and you can figure
out how the various buttons will work even before you begin. Evaluate a
couple of the examples above first to assist you get started.
The objective of python programming is to keep things simple for the
benefit of it users. As should be obvious here, and in the following topics,
there are basic commands that you will have to put forward so as to get the
program working in a particular way. Study these and you can make an
extraordinary program without much stress.

Comments

Python programming language can be used for diverse purposes because it is

one of the most intelligent alternatives that you can give a try when you want
to code. What's more, since it is so simple to utilize, in this chapter we will go
deeper into the study of comments and other unique aspect of python so that
you can be able to start creating amazing codes.
In Python programming a comment is one that will begin with the # sign and
afterward will proceed until you get to the end of the line. For instance:
# This would be a comment
print("Hello, how are you?)
By that comment the system is instructed to print "Hello, how are you?" All
comments are ignored in the Python interpreter since it is all the more a
footnote in the program to support the programmer, or other people who may
utilize the code. They are fundamentally there to state what the program
should do and how it will work. It is more descriptive and can be useful
without affecting how the code functions. Comment is not ment to be in
every line but where needed, on the off chance that the developer feels that
something needs clarified better, they would place in a comment however
don't hope to see it everywhere. Python doesn't support any comment that
will go over a few lines so in the event that you have a more drawn out
comment in the program, make sense of how to separate it into various lines
with the # sign before each part.
Writing and Reading
A few language programs will usually show the content you need on the
screen, or they can demand certain data. You might choose to start
programming by explaining the program to the reader, giving it a name or a
title can make things simpler so the other coder recognizes what is in the
program and can pick the correct one for them.
The most ideal approach to get the correct data to show up is show a
string literals that will add the "print" function. Note, string literals that are
surrounded by quote which can be a single The sort of quote that you use
won't make any difference that much; in any case, in the event that you utilize
one sort to start the statement, you should utilize it till the end. So if there are
double quote at the beginning of your expression, ensure you keep to that
double quote at the end. At the point when you need the system to show a
word or expression on the screen, you would just have "print" and then the
expression after it. For instance, if you need to depict "Welcome!" you would
do
Print("Welcome!")
With this print, welcome appears on your program for others to use, the print
function is going to take up its own line so you will see that after placing this
in, the code will consequently put you on new line. Assuming you want the
guest to do some other activities with your program you can apply similar
principle. For instance, say you need the individual to enter a particular
number with the goal that they can pass through the code, you would use the
string:
first_number = input('put the first number in')
When utilizing the input feature, you won't consequently observe it print on
another line. The content will be put directly after the prompt. You will
likewise need to convert the string into a number for the program to work.
You mustn’t have a particular parameter for this either. In the event that you
do the following option with Parenthesis and nothing inside, you will get a
similar outcome and most times makes it simpler.
Files
Generally, you will utilize the print function to get a string to print to the
screen. This is the default of the print function, yet you can likewise utilize
this equivalent function as a decent method to compose something onto a file.
A genuine case of this is
With open ('myfile.txt', 'w') as myfile:
Print ("Hello!",file=myfile)
This may resemble a straightforward condition, yet there is a considerable
amount of things that are going on in the string over that you should keep an
eye out for. In the spot with you opened up the myfile.txt to compose on and
then assigned it to the variable called myfile. At that point in the next part,
you wrote in Hello! To the file as new line and then the w told the program
that you might have the option to write the changes at the point when the file
is open. Obviously, you don't need to utilize the print function to get it to
accomplish the work that you need. The write technique will frequently
function admirably as well. For instance, you can change the print with
compose/write like the instance bellow to get very similar things.
With open('myfile.txt', 'w') as myfile:
myfile.write("Hello!")
So far we have figured out how to print a strings of words into the program
and even the most effective method to save them to a particular file.
Notwithstanding those options, you can utilize the read method so as to open
a particular file and afterward to peruse the information that is there. In the
event that you might want to open and read a particular document, utilize this
option:

With open('myfile.txt', 'r') as myfile:

data = myfile.read()
with this alternative, you will have the option to inform the program to read
the file contents into variable data. This can make it simpler to open up the
programs that you might want to peruse.
Built in types
Your system is equipped for preparing a great deal of data including numbers
furthermore, characters. The kinds of data that the Python program will
utilize are known as types and the language will contain a wide range of
types to help make things simpler. A portion of these add string, integers, and
floating point numbers. Programmers can even characterize these various
types utilizing classes.

Types will comprise of two separate parts. The first part is a domain
that will contain a potential arrangement of values and the next part is a set
that contains the potential operations. Both of these can be performed on any
value. For instance if you have a domain that is a sort of integer, it can just
contain integers inside it with addition, division, multiplication, and
subtraction. One thing to note with this is Python is a progressively composed
program. This implies that there truly isn't a need to indicate the sorts for the
variables when you make it. Similar factors can be utilized to store the values
of various types. Regardless of this, Python despite everything needs you to
have all the variables with a definitive type. For instance, if the developer
attempted to add a number to a string, the Python program would dictate the
mistake and show error. It won't attempt to make sense of what you needed;
rather it will simply exit.
Integers
On the off chance that you need to utilize integers as a type, you have to keep
them as whole numbers. These can be positive or negative numbers, as long
as there are no decimals with these numbers. In the event that you have a
decimal point in the number, regardless of whether the number is 1.0, you
should utilize it as a floating point number. Python can show these whole
numbers in the "print" function, yet just in the event that it is the sole
argument.
Print (3)
# Let's sum two numbers together
Print (1+2)
On the off chance that you are utilizing whole numbers, you won't have the
option to put the two right close to one another. This is for the most part a
direct result of how Python is a typed language furthermore, won't remember
them in the event that you join them together. If you might want to put the
number and the string together, you have to ensure that the number has
transformed into a string.
Operator Precedence
One thing that you have to monitor when you are working in Python is
operator precedence. For instance, on the off chance that you have 1+2//3
Python could interpret it as (1+2)//3 or 1+ (2//3). Python has a technique that
will assist you with ordering the operation appropriately with the goal that
you get the correct data to appear. For instance, with regards to integer
activity, Python deals first with bracket after which other operations that have
**, then*, and then//, then %, +, lastly - .
Whenever you are composing an expression that has various
operations in it, you will need to remember those signs. This will disclose to
Python how to work the numbers so you can find the correct solutions at that
point. Remember that most arithmetic operators will be left associative so
work it out that way for Python to peruse. The main special case is the **
feature. For instance:
# ** is right-associative
2**3**4
# will be assessed option to left:
2**(3**4)
Strings
While a string may appear to be something complex, in Python they are
fundamentally a sequence of characters. They are going to work a similar
way as a list does, yet they will contain more functionality that is explicit to
the text. Organizing strings can be a test with regards to composing your
code. There are a few messages that won't be fixed string and some of the
time there are values that are stored inside variables within it. There is an
approach to get this to work directly for string formatting. A case of this is:
Name = "Janet"
Age = 24
Print("Hello! My name is %s." % name)
Print("Hello! My name is %s and I am %d years old." % (name, age))
The symbols that have a % first are called placeholders. The variables that go
into these positions will be set after the % in the order where they are set in
the string. In the event that you are doing only a single string, you won't need
a wrapper, however if you do have more than one of these, you have to put
them into a tuple, with a () enclosing it. The placeholder symbols will begin
with various letters, based on the variable type you are utilizing. For instance,
the age will be an integer by the name in a string and all these variables will
be converted into the string before you can be able to add them into the rest.

Escape sequences
Escape sequence is a way used to denote special characters that are usually
difficult to type on your keyboard. What's more, they can be utilized to
represent characters that can be saved for something different. For instance,
using and in the arrangement can confuse the program so you may utilize the
escape sequence to change that for instance:
Print("This is a line. \nThis is another line.")
Triple Quotes
Having dealt with both single and double quotes its important that you
understand that at certain time you may need to bring in the triple quote. This
is used when you have to characterize a literal that will traverse numerous
lines or one that has a great deal of quotes in it. To do this, simply utilize a
single and double together or three singles. A similar principle applies with
the triple quotes likewise with all the others. You should begin and end the
phrase with a similar one.
String Operations
One of the most needful string operations is a concatenation, which is useful
while joining a pair of strings together and it’s denoted with the + symbol.
There are a great deal of functions that Python can support you with and they
will work with the strings to make an assortment of operations. They will
have some helpful alternatives that can do significantly more in the Pythons
program.
In Python programming, strings are called immutable. This implies
that once you make the string, it won’t be changed. You may need to assign a
new valuable to a particular variable that exists in the event that you are
hoping to make a few changes. There is so much that you can find out about
with regards to getting started with Python. It might be a straightforward
language, however you need to have the option to figure out how it functions,
how to write things appropriately, and even how to leave a comment for
others to comprehend when they are glancing through the code. It may appear
to be somewhat scary first and foremost, however sooner rather than later,
and with some training, you will get it down and be composing your own
code within the blink of an eye.

Variables and What They Do in Python

We will be looking straight to the variables; variables are essentially the
marks that points to where things can be stored in your system memory and
the mostly hold values. At the point when it comes to programming that is
typed with statistics, the variables will each have a value that is
predetermined and every variable is just going to hold the value of that type.
Python has made it somewhat simpler in light of the fact that you can utilize
one of your variables so as to store various types. Consider your calculator
for instance. The variable will resemble the memory function in this
calculating machine. It will hold onto a value with the goal that you can
recover it whenever that you need to, however when you store in a more
current value, the old ones one will be deleted. The main difference is that
you will have the option to have a large number of variables and every one of
them will have various values, each of them being alluded by their own
name. With Python you will have the option to characterize a variable by
giving the mark a value. For instance, you can name a variable count and
have it a integer value of one. You would show this by basically writing
count = 1
Note: that with this synthax, you can allot a value to the variable with the
equivalent name. On the off chance that you attempt to access values in a
variable that hasn't been characterized, the Python interpreter won't read
through this. It will simply exit out of the program what's more, give you an
error. You can as well decide to characterize a couple of various variables in
a single line, however this isn't the best practice for you to utilize. For
instance, you could do this:
# Let's characterize three variables simultaneously:
count, result, total = 0, 0, 0
Though that is the right method it’s best to display it like this:
# This is same as:
Count = 0
Result = 0
Total = 0
It is a lot simpler to peruse the subsequent way and will guarantee that the
Python program will comprehend what you need it to state.

Understanding the scope of a variable

You won't have the option to get to each variable from all parts of the
program and Not all the variables only one out of every odd variable will be a
similar length. The way that you characterized the variable will determine
where and to what extent you will have the option to access this variable. The
area of your program where you can get to the variable is going to be known
as the "scope" and the time that the variable will be accessible is known as
the "lifetime". Global variables: are those that are characterized inside the
basic file body and you will have the option to see these variables all through
the whole file as well as within a file that can import the particular file. These
variables have far reaching impacts and along these lines, you may see a few
results that you didn't take note. This is the reason a great many people won't
utilize global variables, or they will utilize them sparingly. You should
possibly include stuff into the global namespace if you intend to utilize them
comprehensively, as with functions or classes.
But on the off chance that you characterize a variable within another
variable, it will be called a local variable. This one can be accessed from
where it is characterized and will possibly exist when that function executes.
These are just going to be accessible in specific areas of the program and
can't be found or utilized somewhere else.
The assignment operator
We have talked about this option a bit all through the book, yet haven't
generally given it a name. The assignment operator is represented by the (=).
It is going to be utilized in programming to assign the value to right of the
statement to the variable that is found to the left. Sometimes the variable will
be made first. In situations where the value on the right is from an expression,
for example, a number arithmetic expression, the evaluation will happen
before this assignment occurs.
Remember that the (=) won't be a mathematical sign in programming.
You can add things to the number and make a wide range of changes that
wouldn't make sence on the off chance that you thought of this sign as a
mathemtically one. Rather it is an assignment administrator with the goal that
the statement will be transformed into the part on the right. At the point when
you choose the first value to this variable, you are experiencing the procedure
of initializing. The characterization of a value assignment and variable are
completed in the single step in this programming, in spite of the fact that it is
here and there done in two steps with a portion of the other programming
languages. Be that as it may, since it is done in one step, it is more outlandish
that the user will commit an error or get an error during the process.

Modifying values
In some programming languages, you will have the option to characterize a
unique variable that has a value that has been set. This implies the value can't
be changed. These are called constants in the programming language.
Generally, Python won't take into consideration these sorts of limitations, yet
there is a convention that is utilized to help guarantee that a few factors are
set apart to demonstrate that the qualities should be changed. To show this,
the names will be written in CAPITAL letters with underscores between each
word. An example of a variable that is a constants include:
NUMBER_OF_DAYS_IN_A_WEEK=7
NUMBER_OF_WEEKS_IN_A_YEAR=52
Obviously, there are no rules to state you need to put the correct number
toward the end. You could state there are 8 days in a week if you want it that
way on the grounds that the Python program won't follow along, however it
is ideal to simply keep it exact in the event that other coders might want to
utilize it. These can be extremely useful to you in your string. Some of the
time in the program, for instance, you will need to change the limit of a
number that is permitted in the program. This may work fine for a bit, yet
perhaps later on you have to increase or reduce this number. Without setting
up a constant, you need to experience and make many changes to get
everything coordinated. Yet, with a decent constants all together, you can
simply return to one spot and get everything repaired. Understanding how the
strings work in your program can have a big effect in the success that you see
with this program. You have to realize where they are stored, what the rules
are that administer every one of them, and how to make them work in a
particular part of the program. With a touch of training, and utilizing the
procedures above, you will get this down in no time and can be a veteran as
well!

Python programming
2

How does Python Execute a Program

 Each programming language that you work with will execute a
program in an unexpected way. This is the reason it is so imperative to figure
out how to sort out the words and the various statements that the program
requires so you maintain a strategic distance from mistakes and different
issues in your code. In this section, we will set aside some effort to figure out
how Python will execute the orders that you give and essentially how the
entire program runs. At the point when you are working with Python, you are
working with a interpreted programming language. You will have a text
translator that will execute every one of the programs going line by line and
afterward will change it into a code for the process to comprehend the words
what's more, complete them for you. Python is additionally a scripting
language, so you can work out the content and afterward save it utilizing the
extension.py or you can straightforwardly write it and afterward execute
every statement into the Python shell.
Internally, Python is going to work to accumulate your program,
fundamentally the source code, into a byte code that has the .pyc extension,
much the same as the Java byte code. This makes it simpler for the code to be
executed without any delays and you will have the option to see it come up in
only a couple of moments rather than waiting around. You will have the
option to save your byte code records into a subdirectory that is called
pycache which is situated in the directly where the source file dwells. For
instance, in the event that you worked out helloworld.py it is going to, at that
point be changed over into one of these byte codes and renamed
helloworld.pyc. You can go in and manually compile this code if something
turns out badly, yet for the most part, Python will do the compilation for you
so it won't be an issues. As a starter you may wonder where some of these
pyc suffix originate from, yet Python is going to store them with that
particular suffixe so whenever it shows up don’t feel threatened. Obviously,
this is possibly going to occur if Python has the write access, though even if
the python has no write access, it may not be saved that way, but the program
will still work. At whatever point you call up a Python program, Python is
going to check if there is an existing compiled version with this .pyc suffix.
This file ought to be more current than the .py suffix and on the off chance
that it exists, the Python will load in the byte code to accelerate how quick
the script is capable to go. If the byte code doesn't exist on your system,
Python will work to make your byte code before it executes the program. So
essentially, each time that you execute a script in Python, you will have a
byte code made by the program also. On the off chance that the script in
Python is imported like a module, your byte code will be stored in the best
possible .pyc file.
Python Implementations
At the point when you hear about implementation of Python, it implies that
the program or the environment that is offering support for executing your
programs within the Python language, will be represented with the CPython
reference implementation. This implies that it will assist you to work on
executing the various codes and statements that you are taking a shot at
within the program. There are likewise a few variations of the CPython that
you can take a shot at and will have a major effect in the manner that the
program works. Some of the features that are availiable with the variations
include:

1. CrossTwine Linker—this will be a mix between CPython and an

add-on library of your choice. It will offer some better presentation
when it goes to the code that you are working on.
2. Stackless Python—this is CPython that has an accentuation on
concurency while utilizing channels and tasklets. This is regularly
the benevolent that is utilized for the dspython on programs like
the Nintendo DS.
3. Wypthon—this is viewed as a re-implementation of some of the
parts of Python, which will drop the support of utilizing bytecode
so as to utilize the wordcode based omdel. It will utilize the stack
register in the implementation and includes in lots of different
types of optimization.

implementation is everything with regards to how you can work at your

programs and can assist you with getting progressively finished with Python
contrasted with some of the other programming languages. What is so
exceptional about Python is that it can work with some of other programming
languages so as to maintain its versatility and retain the power to get things
done easily.Some of the other implementations that you might need to
consider on the off chance that you need to accomplish something explicit
with the Python programming language include:

Brython—this is one of the implementations that you can use so as

to run Python in your program utilizing a translation to JavaScript
so you can use these two together.
CLPython—this is an implemenation of Python in similar lisp.
 HotPy—this is viewed as a virtual machine for Python that will
bolster translation and bytecode improvement.
IronPython—this is Python in C#. C# is an incredible
programming language to utilize within the Windows stages and is
frequently a contender to Python dependent on how well known
and simple it is to utilize. This implementation permits you to
translate your work from Python over to C# in the event that you
pick.
Jython—this is the version of Python accessible for the Java
system.
PyMite—this is Python that you can use for inserted gadgets
PyPy—this is Python inside Python with the goal that you can
focus on a couple of various environments simultaneously.
RapydScript—this is a language that is like Python that will
arrange into
JavaScript so you can utilize it in the Java stage without going
with all of the troublesome language issues.
Taking a shot at Python can be an incredible encounter. On the off chance
that you are only a learner with the possibility of programming and are
uncertain about how to kick the entirety of this off, some of the other
programming dialects can be somewhat confounding. Python is anything but
difficult to utilize however has all the power that you need from a portion of
the greater names in programming language and you get the advantage of
getting the opportunity to utilize this program alongside a portion of the other
well known languages that you might need to work with!

Survey of a Simple Program

As we referenced, utilizing Python is one of the easiest programming
languages that you can work with. It isn’t complex like other languagesIt
additionally makes it simpler for you and for another person to go through the
information and have the option to peruse it. So how about we take a look at
some of the things you can do with Python and how to begin with composing
your first program.
The primary program that we will start with is the "Hello World" program.
This one is going to require a Python shell to make it simpler and you will
have the option to test it out on your editors on the off chance that you do it
appropriately. This makes it simpler to have a smart thought of what you are
doing and to detect any errors at the outset. In the event that you are utilizing
the Python Shell, which functions admirably on the vast majority of the PC
types and programs that you might be utilizing with Python, you will
basically need to type in the following programs to get the information to
appear:
Print("Hi World!)
You ought to have the option to proceed to execute this data and find that it
will appear with the words Hi World! On the screen. This is a basic
procedure to do, however it will help you to kick things off and gives a
decent review of some basic steps that you need so as to begin composing
your own program on Python. Don’t forget that Python is an extremely
simple programming language that won't have a ton of various brackets and
other data that is standing out that can slow what you are doing. It is
additionally extremely simple to peruse.
As should be obvious, you only need a couple of things set up so as to
work out the phrase, instead of expecting to compose lines of code to get a
similar outcome like you would need to do with other programming
languages. We should look into some of the different things that you can do
with Python programming and how you can even think of some great code to
kick your programs runing.

Some of the Basic Commands, Variables, Statements, and Other Things

that you can do
There are such a significant number of things that you can do so as to get a
code fully operational on Python. Numerous individuals may abstain from
utilizing Python since they feel that it is excessively basic or it simply won't
take care of business. However, in realworld, it is basic only for a beginner to
learn how to use it. But that does not mean that you wont be able to how tithe
way that even a tenderfoot can figure out how to utilize it, yet that doesn't
imply that you can't do indept work using python This section is going to set
aside some effort to take a look at the various orders that you can do with
Python programming so as to make your projects and codes wake up.
Variables
Variables when mentioned in programming language sound strange to the
understanding, but they are simply locations in the memory that are reserved
for storing the values of your code. At the point when you take a shot at
making a variable, you are saving this spot in the memory. Sometimes, the
data type that is in the variable will advise the interpreter to save the memory
space and can even choose what you can store on your reserved memory.

Allocating values to your variables

The value will be one of the fundamental things that your program should
work with. it very well may be a string, for example, Hi World, 3.14, which
is viewed as a type float, or an integer like 1, 2, 3. Python variables won't
need an explicit declaration so as to reserve the space in the memory that you
need. This is something that will happen consequently at whatever point you
place a value with the variable. For this to work, just put the (=) with the goal
that the value knows where it is expected to go.
A few instances of this include:
X = 10 #an number assignment
Y= 200 #an integer assignment
Pi (π) = 3.14 #a floating point assignment
Empname = "Arun Baruah" #a string assignment
Remember that when you are working on codes, you can leave a comment
with your wok by utilizing the # sign. This permits you to clarify what is
happening in the code, leave a few notes, or have something different inside
the program. It won't be perused by the translator since it is only a little note
that you kept for yourself or for another person comprehension. The
following part will rely upon which form of Python you are utilizing. Python
2 is fine with you writing out print and afterward the data you need to discuss
but Python 3 will expect you to put the parenthesis in to make it work. For
instance:
Print("y = %d" %y)
Print("x = %d" %x)
Print("Employee Name is %s" %empname)
These would then pass through the interpreter and the results you would get
ought to be
X = 10
Y = 200
Employee Name is Arun Baruah
Now attempt running this program and observe what comes out. On the off
chance that you didn't get the results as listed above you should cross check
to confirm were the error is coming from. This is an easier way try out your
abilities on python programming.

Multiple assignments
Notwithstanding working with the single variables that were listed above,
you will likewise be ready to work on multiple assignments. This implies you
will be ready to assign one value to a few distinct variables simultaneously.
To do this, you would simply require to put the equivalent (=) sign between
every one of them to keep things sorted out and to tell the system that the
value will be with the whole of the factors together. You can keep them
separated out if that is best for you, however utilizing this technique is going
to assist you to send everything to a similar memory location on the system
and will give the code a more clear look on your screen. An instance of how
to give more than one variable a similar value includes:
a=b=c=1
From the above you want to let the code know that all the variables are of the
same value and that you want all the variables to be at same location within
your memory.
Standard data types
In addition to the goodies of python another thing that you’re able to work is
the various data types which will be used in your code to define the
operations that you can do on each data type and explain to others the storage
method that will be best for this kind of data. Python has five standard data
types, they are:
Numbers: The Number data types are the ones that will store the numeric
values. They will be made as objects once you assign a value to them. There
are additionally four unique kinds of numbers that Python will support they
are :

Complex, (for example, complex numbers)

Float (floating point real values)
Long (Long integers that can likewise be seen as hexadecimal and
octal.)
Int (signed integers)
One thing to note is that while Python will permit you to utilize the lowercase
l while doing the long type of a number, it is ideal to go with a capitalized L
at whatever point you are utilizing the letter. This is going to keep you away
from confusion while reading the program between the l and, the 1 as they
look extremely comparable. Whenever that Python is showing a long
integerthat has the l in it, you will see the capitalized L.
Tuples
The tuples share some similarities with the lists but the difference is that
tuples uses different signs. The primary difference however is that lists will
utilize brackets and the components, just as the size, can be changed through
the program. Although, the tuples are going to utilize brackets and you won't
have the option to update them. A decent method to consider tuples is that
they will resemble a read only page. Once you don't attempt to make changes
to the tuple in the program, you will be able to utilize it similarly as you did
the list examples above. This makes it a good option to utilize it in case
you're interested on something that is straightforward yet won't let anybody
make changes to the program after you are finished.
Lists
Lists are one of the most flexible data types that you can take a shot at in
Python. In this language, the list will contain various things that are either
enclosed with the square brackets or separated out with commas. They are
like the clusters that you would find in C on the off chance that you've
worked with that program. The one distinction that surfaces with these is that
the items that are in a list can be from various data types. The values that are
put away inside the list can be gotten to with a slice operator also as the [:}
image with the indexes beginning at 0 toward the start of the list and
afterward going down until you get to - 1. The plus sign will be the
concatenation operator while you can utilize the asterisk as the repetition
operator. For certain instances of what this means and how you can utilize the
various signs inside your programming, think about a few of these models:
list = ['mainu', 'shainu', 86, 3.14, 50.2]
tinylist = [123, 'arun']
print)list)#prints complete list
print(list[0]) #prints the first component of the list
print(list[1:3]-#prints components beginning from the second component and
to the third
print(list [2:]) #prints the whole of the components of the list beginning with
the third component.
Print(tinylist*2) #prints the list twice.
Print(list + tinylist) #prints the concatenated lists.
Strings: Strings are distinguished in Python as a contiguous set of characters
that will be displayed by the utilization of quotations marks. Python will
permit either double quote or single quote, however you do need to keep
things sorted out. This implies in the event that you utilize a double quote
toward the start of your string, you have to end that equivalent string with the
double quote. The equivalent goes when you are utilizing a single quote.
Both of these will mean the same thing, you simply need to ensure that you
are utilizing the best possible quote marks to make the code look great and to
abstain from making the Python program look complex.
Notwithstanding having the option to print off the string that you might
want, you are also allowed to instruct the program to print simply part of the
string utilizing some exceptional characters. How about we look at some of
the examples of what you can do with the strings, and the corresponding
signs that you will use as well, to help illustrate this point.
str = hi Python!'
print(str) #prints complete string
print(str[0]) #prints the first character of the string
print(str[2:5]) #prints characters beginning from the third to the fifth
print(str[2:]) #prints string beginning from the third character
print(str*2) #prints the string two times
print(str+"Guys") #prints concatenated string
Generally you are most likely going to need to print out the whole string to
leave a message up on your program so the first print that you do will be
sufficient. In any case, if you simply need to print out Hi or some other
variation of the words above, you may discover that different options are
extremely valuable. You can do any combination of these, they are only
guides to assist you as a first timer!
Dictionary
Dictionaries are another sort of hardware that you can utilize when you are
working in Python. They are like a hash table type and they are going to work
like the hashes or the arrays that you can discover on other programming
languages like C# and Perl. They will likewise comprise of key value pairs
and keeping in mind that the key can be practically any type on Python, you
will see that they are typically going to be strings or numbers. For the most
part, with regards to values, you will find that they are an arbitrary objects in
python. A few instances of how this will function include the accompanying
codes:
#dictionary stores key-value pair, later to be retrieved by the values with the
keys
dict = {}
dict['mainu'] = "This is mainu"
dict[10] = 'This is number 10"
empdict = {'name': 'arun', 'code':23, 'dept': 'IT'}
print(dict['mainu']) #this will print the value for the 'mainu' key
print(dict[10]) #this will print the incentive for the 10 key
print(empdict) #this will print the complete dictionary
print(empdict.keys()) #this will print out the whole of the keys
print(empdict.values()) #this will print all the values
One thing to remember is that these dictionary values won't be stored in an
organized manner t. They won't have the concept of ordering among the
components. This doesn't imply that you can say that the components are out
of order, they are simply going to be unordered.

Keywords
The majority of the kinds of programming languges that you will manage
will have a few keywords that are saved as a component of the language.
These are words that you truly shouldn't use in your code except if you totally
can't resist. There are 33 keywords found in the latest version of Python and
you should spell them appropriately on the off chance that you need them to
carry out the responsibility that you want them do. The 33 keywords that you
should keep an eye on include:
False
Class
Finally
Is
Return
None
Continue
For
Lambda
Try
True
Def
From
Nonlocal
While
And
Del
Global
Not
Yield
As
Elif
If
Or
Assert
Else
Import
Pass
Break
Except
In
Raise
Keep this list handy in the event that you are stressed over learning the
language. It will have the option to help you out whenever that you have
issues with the interpreter about the names that you are giving the variable.
You might be confounded regarding why it is giving you a few issues with
the words you picked, you can through this list and check whether you
utilized one of the keywords inappropriately within your code.
Statements
At the point when you are composing your code in the Python language, you
will be making expressions and statements to get the program working.
Expressions will be able to process the objects and you will discover them
installed within your statements. A statement is essentially a unit of code that
will be sent to the interpreter with the goal that it may be executed. There are
two sorts of statements that you can utilize; assignment so far and print. You
will have the option to compose the statement, or multiple statements,
utilizing the Python Shell to do so intuitively or with the Python script
utilizing the .py extension. At the point when you type these statements into
the interactive mode, the interpreter will work to execute it, as long as
everything is appropriately set up, and afterward you can see the outcomes
shown on the screen. When there are many lines that you have to write in
code, it is ideal to utilize a script that has an arrangement of statements. For
instance
#All of these are statements
X = 56
Name = "Mainu"
Z = float(X)
Print(X)
Print(Name)
Print(Z)

Operands and operators

There are a ton of incredible symbols that come upwhen you make a code in
your Python program. It is imperative to comprehend what parts you can
work with and what they stand for. Operators are regularly used to mean
subtraction, addition, division, and multiplication. The values of the operator
will be called operands. You can utilize various signs for these, so as to get
the values that you might want to see.

While you are utilizing the operators and operands, you have to recall that
there is

going to be an order of evaluation depending on the rules of precedence.

While working on arithmetics Python is going to utilize the abbreviation
PEMDAS which is bracket, exponentiation, parenthesis, exponential,
multiplication, and division. On the off chance that there are some of these
that are the same, for example, two arrangements of numbers that should be
multiplied together, you will require to work from left to right to get the
accurate number. Another significant operator that you should search for is
the modulus operator. This one is going to work with whole numbers and is
going to yield the remainder once the first operand has been divided by the
second one.

Note: one other good thing about python is that you can send out, which
permits you to explain what you are working on instead of stressing the other
programmer that will read the code to comprehend the code that way. You
will simply introduce the # sign to denote that a comment/statement about the
code was left. The program interpreter whenever it comes across the # sign
will ignore it but it won’t be removed because another programmer may need
to look at the code if need be.

Understanding the decision control structure

Human beings at some times will have to make decisions due to the situation
around them, assuming you wake up one morning nd wanted to go for work,
but it began to rain out. Did you simply stay there confused without another
alternative to support you out? No, you may have woken up and chosen to
take an umbrella and go on the walk anyway or you may choose to remain at
home and do something else. You can settle on different choices, regardless
of whether the first doesn't work out, in view of the conditions around you.
Presently this is somewhat a similar thought with regards to working with
Python. So far we have just advised the program to do each thing in turn. On
the off chance that the circumstances don’t arrange precisely with the
program, you won't get any outcomes. This won't work for a few programs,
particularly if the others are permitted to pick from several unique answers.
The decision control structure is the part that permits you to pick a couple of
various options for Python on the off chance that the first decision doesn't
work out.
Generally, these are going to work at a true or false kind of outcome.
You will need to make sense of which action you need to make and what
statement the program ought to execute if the result is either true or false. In
Python, any answer that is non-null or non-zero will be viewed as true and
the ones that are either invalid or zero will be seen as false. For more
understanding of these consider the following: To see a portion of these think
about the accompanying:

If statements: the if statement is going to comprise of a Boolean

expression that is at that point followed by at least one statement
that will be executed if the appropriate response coordinates.
If… else statements: this option will have a statement that will
appear if the "if" statement is right, however there is likewise
another statement that is permitted to come up in the event that the
Boolean statement ends up being false.
Nested if statement: you can utilize one if or if else statement
within another one when required.
A few instances of how this functions incorporates:
age = 23
if (age == 23):
print("The age is 23")
print("Have a good day!")
At the point when this code comes through it will state:
The age is 23
Have a good day!

Utilizing the "if" keyword is going to tell the compiler that what you are
composing is a Decision control instruction. The condition that is behind the
keyword if it is within the parenthesis, the conditions should be true in the
event that you need the code to come out, however on the off chance that it
isn't correct, this statement is simply going to be overlooked in this
circumstance and it is going to proceed onward to the next command that you
give. You can as well set up an "if… else" clause. This one would show the
unique message if the conditions were true. In any case, if the conditions are
false, it would end up putting out a second statement. This can assist with
guaranteeing that you are getting the correct message across regardless of
what another person is sending over and keep the interpreter from totally
overlooking this step.
 So the following inquiry that you might be asking is the means by
which we can tell whether the statement is true or false? You should utilize a
portion of the relational operators to make this happen on the grounds that it
will have the option to compare the two values to check whether they are
equivalent, inconsistent, or another choice. A portion of the options that you
can use to check whether a statement is correct include:
Expression Condition for the expression to be
true
X == y X is equal to y
X != y X X is not equal to y
X<yX X is less than y
X>yX X is greater than y
X <= y X X is less than or equal to y
X >= y X X is greater than or equal to y

Below is a good example

age = int(input("Enter your age:"))

if (age <=18):
print("You are not eligible for voting, try next election!")
print("Program closes")
This code will print you out a difference example based of the age that is
placed in. Since it is an if statement in particular, you will find a solution
exactly when the number is 18 or under. Here we will take a look at the
output based on the age of 18 and that of 35.
Enter your age: 18
You are not qualified for voting, try next election!
Program ends
Enter your age: 35
Program closes
Presently we should investigate including various statements in your "if"
statement. It isn't extraordinary to discover at least two statement being put
inside an expression and working fine and dandy if everything is fulfilled. On
the off chance that these statements are executed, you will need to ensure that
you indent them appropriately. We should investigate how this could work.
Ensure to check out this on your interpreter to get some knowledge in
composing code and how the "if" statements are going to work.
bonus = 0.0
currentyear = int(input("Enter current year:"))
yearofjoining = int(input("Enter year of joining:"))
yearofservice = currentyear – yearofjoining
if(yearofservice >2):
reward = 1500
print("Bonus - %d" %bonus)
print("Congratulations! We can give you a reward o %d" %bonus).
Now if the "if" statement ends up being higher than two, you will find that
the congrats statement is going to be displayed. However, in the event that
the sum is lower than two, you are going to wind up without any statement
since the entirety of the statements were not met. You can put whatever
numbers that you need inside it so as to get it to work for each employee in
this option.
The "if-else" Statement
So far we have recently been discussing the "if" statements. These are ones
that should be true before any of your statement come out. On the off chance
that things come out to be false, there will be no statement by any means.
Now assuming you want to have different statements displayed then you will
need the "if else" statement, you can pick two or more, statements that are
going to come up dependent on the outcomes. On the off chance that your
outcomes end up being true, you will have the first statement appear,
however on the off chance that the outcomes wind up being false, you can
pick another statement that you might want to appear also. This guarantees
you are getting an answer regardless of the outcome with the goal that the
program shows something new. You simply need to ensure that subsequent to
working out your "if" statement, you include the "else" and afterward put in
the statement that you might want to have displayed. This can take a touch of
time to work, yet it opens up such a significant number of extraordinary
options with your code to ensure that everything levels out and looks pleasant
with your code.
The elif Statement
Another alternative that you can do with your statements is the elif statement.
This one is going to give you the choice of looking at a couple of expressions
as true, instead of just one expression as true, with the goal that you can
execute the entire block of code once just one of the conditions goes up to be
true. Obviously, doing this is discretionary, yet there is the advantage of
being able to have any number of elif statements after the if.
How about we investigate what the whole of this is going to look like when
you work it out in the syntax:
if expression1:
statement(s)
elif expression2:
statement(s)
elif expression3:
statement(s)
else:
statement(s)
You will at that point have the option to put your data into the parts and find
the solution that is

listed with every one of the parts. For example:

Print("Let's enjoy a Pizza! Alright, let’s go inside Pizzahut!")
print("Waiter, Please select Pizza of your choice from the menu")
pizzachoice = int(input("Please enter your choice of Pizza:"))
if pizzachoice == 1:
print('I want to enjoy a pizza napoletana')
elif pizzachoice == 2:
print('I want to enjoy a pizza rustica')
elif pizzachoice == 3:
print('I want to enjoy a pizza capricciosa')
else:
print("Sorry, I don’t need any of the listed pizza's, it would be ideal if you
bring a Coca Cola for me.")
you will see the data that is listed after print appear for the initial three lines.
After the third line, the program ought to ask you to place in one of the four
options. Depending on which option you pick, you will be able to see the
correct answer, either the pizza of your choice or the option to simply get a
beverage. It might resemble somewhat of a wreck from the start, yet it causes
you to get alternatives that the code will have the option to comprehend while
restoring the right "pizza type" to you all the while.
The "if" statements will give you a great deal of help when you are
looking to make some extraordinary things occur in your code, you can
likewise utilize nested if statements that lets you to include a couple of these
inside one another, adding some more power to your code. Obviously, it
might take some understanding to get this all down, yet you are going to be
amazed at all that you can do when you are chipping away at your codes.

Loop Control Statements

Up until now, we have talked about a great deal of programs and what you
can do when you are working in Python, yet these are on the whole going to
be either choice or sequencial control instruction. For the initial ones, we
were doing calculations that will be done in afixed order while with the
subsequent one, the correct set of instruction were executed based on the
result of the conditions that were tried. There were a few confinements as a
result of the manner in which they are executed and they are as it were able to
play out precisely the same series of action, consistently similarly, and they
are just ready to do it one time.
There are times when you will need to work out a code that can be
more entangled. One of these options is for the loop statement. This sort of
statement will permit the programmer to execute a statement, or even a batch
of statement, a few times. On the off chance that you have a statement that
you might want to keep returning in the program, you will need to make your
own loop statement to get this going. It is conceivable to utilize Python so as
to hand these loop statements and there are three strategies that you can pick
so as to cause the loop statement to occur. These three techniques include:

Nesting loops
Using a for loop
Using a while loop

The While Loop

The first sort of loop that we will take a look at is the while loop. This is a
decent one when you want the code to accomplish something for a fixed
number of times. You would prefer not to have it go on uncertainly, however
you would like to have it go for a certain amount of times, for example, eight
times, before stopping. Study the example below for more knowledge on how
the while loop works.
#calculation of simple interest. Request that user input principal, rate, number
of years.
counter = 1
while(counter <= 3):
principal = int(input("Enter the principal amount:"))
numberofyeras = int(input("Enter the number of years:"))
rateofinterest = float(input("Enter the rate:"))
simpleinterest = principal * numberofyears * rate/100
print("Simple interest = %.2f" %simpleinterest)
#increase the counter by 1
counter = counter + 1
print("You have calculated simple interest for 3 time!")
The result is going to come out to permit the user to put in the data that they
want to compute. It will make sense of the interest rate and the amount that
will be determined dependent on the numbers that you give. It will go on a
loop with the goal that they are able to do this multiple times for utilization of
this. You can set it up to take on more if you would like, however for this one
we are simply utilizing it 3 times for straightforwardness.
The "For" Loop
This is for loops that you will utilize when you need a piece of code to rehash
a certain amount of times. It is somewhat not quite the same as the one above
in light of the fact that it is a greater amount of the conventional approach to
get things done, however it can in any case be valuable. This alternative will
be a piece not the same as what you will discover in C++ and C.
Instead of this loop allowing the user to characterize the stopping
condition or the iteration step, Python will have the statement iterate over the
items in the order that they appear in the statement. Below is an instance:
# Measure some strings:
words = ['apple', 'mango', 'banana', 'orange']
for w in words:
print(w, len(w))
At the point when it experiences the loop with this, you will get the four fruit
words above turn out in the order that they are composed. On the off chance
that you need them to be done in an alternate order, you should submit them
in an alternate order when you are putting them into the syntax, to keep
things simple. You won't have the option to make changes to the words when
they are in the syntax like above.
Assuming you need to iterate a little more than a particular sequence of
numbers, utilizing the Function range() can truly prove to be useful with this.
It will create an entire list containing some of the number arithmetic
progressions that you are hoping to utilize.
Nested loop
This is fundamentally just one loop that is within another and it will keep
running until both of the programs are finished. This can be helpful for
various things that you need to do with your program, however below is an
example that will give out the multiplication table going from 1 to 10:

#write a multiplication table from 1 to 10

For x in xrange(1, 11):
For y in xrange(1, 11):
Print '%d = %d' % (x, y, x*x)
when you get the result of this program, it will appear to be like this:
1*1 = 1
1*2 = 2
1*3 = 3
1*4 = 4
As far as possible up to 1*10 = 2
At that point it would proceed onward to do the table by twos, for example,
this:
2*1 =2
2*2 = 4
Etc until you end up with 10*10 = 100 as your last spot in the sequence.
These loops can assist you with getting various things to appear on your
system, sometimes uncertainly, however they are going to prop up through
the loop just for the amount of times that you might want it to. You will have
the option to put these loops in with a touch of training and there are such a
large number of things that you can do. For instance, the simple formula
above will give you the whole multiplication table beginning with 1*1 and
finishing with 10*10. For such a straightforward statement, you are getting an
incredible measure of data from it and a significant number of the loops that
you work with will be the same.

Functions
Functions are another significant part of learning the Python language. These
are fundamentally blocks of code that are independent and will have the
option to play out some sort of rational task in your code. At the point when
you set aside the effort to characterize a function, you will be able to specify
the name of the function just as the sequence of the statement. You will at
that point have the option to call up the function utilizing its name. There are
two kinds of functions that we are going to use here and we will talk about
them beneath.
User characterized capacities
With this one, you will have the option to characterize the function that you
are utilizing. They are going to have essentially similar guidelines that you
found with variable names. This implies using underscore characters,
numbers, and letters will work extraordinary, however you ought to never
utilize a number for the first character in the function. You can't utilize your
keywords as the name of your function and you ought to be cautious about
having a function and a variable that have a similar name. The empty
brackets that you will see after the name will show that the function isn't
going to take on any arguments. The first line will be known as the header
while the rest are known as the body. You have to ensure that the header
closes with a colon and that you indent in the body so the interpreter
recognizes what you are doing. A good instance of the syntax or a function is:
def functionname(arg1, arg2, arg3):
'"docstring of the function i.e., a short presentation about the function"
#code inside the function
Return[value]
#a function might not have any arguments or parameters like
def functionname():
#code inside the function

Utilizing the parameters of a function can be substantial in any of the data

types that you are utilizing with Python, regardless of whether you are
managing user characterized classes, dictionaries, tuple, list, float, and int.
We should investigate a portion of the various pieces of this function so you
can comprehend the significance of every one and how they work.
Docstring function
. The first string that you see directly after the header in a function is known
as the docstring, something that is short for documentation string. It will be
utilized in the code so as to clarify what the function does. This is a
discretionary part, however it is a decent practice to get into. In the event that
you are anticipating having this go over a couple of lines, you ought to
consider doing the triple quotes so as to let the system know what you are
doing.
The return statement
The function is continually going to give back a value. The return statement
will be utilized as a exist function and will return to where it was called from.
This statement is permitted to contain expressions that it can evaluate before
giving out a value. If there are no expression in the statement, or the return
statement isn't even present within the function, you will find that the return
you get is the None object.

Lifetime and scope variables

Something else that you can find out about the Python language is about
lifetime and scope variable. The scope variable is the part that will be seen
and obvious. The variables and the parameters that are in the capacity won't
be obvious from the outside eye (the individuals who are taking a look at the
code when it is being executed) yet they will have a local scope that will be
displayed.
On the other hand, a lifetime is the timeframe that the variable is going
to exist in the memory. The lifetime of most variables within functions will
be as long as your function executes. After the function returns the value,
these will be decimated with the goal that you can place in no information
sources and get various outcomes. For instance, in the event that you input
certain numbers and, your return winds up being five, next time you’re able
to put a different information and you may get a six. These will erase after
each function has gone through with the goal that you can get new outcomes
if new data is entered.

Pass by references
In Python, the entirety of the parameter that you put in will be passed by
reference. This implies that the location of the area in the memory will be
referenced to the memory area. So in the event that you are going to change
what your parameter is able to refer to in the function, this equivalent change
is then going to reflect back when you are doing the calling function.
Flow of execution
The execution of the statement is continually going to begin directly at the
first statement that is in the program. Your statement will be done just one at
a time to stay away from disarray and ensure that the program is running as
easily as it should. The execution is likewise going to happen going from top-
down so you have to ensure that you are placing them rightly. Ensure that
you are characterizing the function before you choose to ring them to get the
correct request.
Anonymous functions
Python permits the developer to make anonymous functions. These are
functions that won't be bound to a name at run time and you are going to need
to utilize a construct that is known as "lambda." This operator is an approach
to create these functions that try not to have a name. Fundamentally you will
need to make these when the functions are considered expendable, or they are
simply required right where they have been made. There are a couple of
functions that Lambda functions will work with including reduce(), map(),
and filter().
The syntax that you will need to use with the lambda function is:
lambda argument_list: expression.
The reduce, filter, and map function
We have discussed various functions so far in this book, yet it an opportunity
to take a look at a Few others that are going to help you with get more out of
the code that you are composing. Especially, we are going to study the map,
filter, and reduce functions, with list comprehension to assist you getting
started.

Map function
The map() function is the one that will apply to every part in an iterable, or
those within a list. Generally, you would utilize the anonymous inline
function to make this work, however this is one of them that you can use
within any of your functions. So if you are attempting to take a shot at a list,
this would be a decent one for you to utilize in your system code.

Filter function
Following up is the filter function. Much the same as you would figure from
the name, the filter can remove the components in the sequence for which the
functions returns a true. The rest would be overlooked. This implies you will
need to do a grouping that would be either true or false so as to get this one to
function. At the point when the sequence has a couple of alternatives that are
valid, the filter function will choose those ones. Though, if every one of
them, or possibly a portion of the sequence points, are false, you will find that
these are not separated out.

The reduce function

This one is somewhat special. It will take a few values to the sequence and
will work it until you end up with only one value, as opposed to the
enormous list that you have. It is going to work from left to right so as to find
the solutions that you are searching for. Depending on the length of your
sequence, you may need to accomplish some work with this to get everything
done.

We should take a look at the numbers 1, 2, 3, and 4. The reduce function is

going to take these four numbers and transform them into one. It will do this
by gathering the 1 and the 2 into a single unit to get 3, at that point gathering
the 3 and the 3 into a single unit to get six, and afterward gathering the 6 and
the 4 into a single unit to get 10. So the answer will end in a single value of
10. A decent syntax to show how this works is
from functools import lessen
results = reduce( (lambda x, y: x +y), [1, 2, 3, 4])
print(result)
List comprehension
This is an extraordinary method to make a few lists within Python. A portion
of the normal applications will utilize your components to make the new list
and others will make a subsequence of these components when they satisfy
certain conditions. The list comprehension feature can really be a substitute to
utilizing the lambda function also as the other functions that we just
discussed. This is on the grounds that the list function is simpler to work with
and comprehend. The syntax of utilizing the list comprehension is
[expression-involving-loop-variable for loop-variable in sequence]
This is going to step right over all the components of the sequence so you can
set up the loop variable for every component each in turn and afterward it
totally disposes of what you required the lambda function for in the first
place.
Conclusion
Working in Python can be a standout amongst other programming languages
for you to pick. It is easy to use for even the computer illiterate, yet it has the
right force behind it to make it an incredible programming language
regardless your programming level or skill. There are simply such a large
number of things that you can do with the Python program, and since you can
blend it in with a portion of the other programming languages, there is nearly
nothing that you can't do with Python on your side. It's anything but an issue
in the event that you are truly constrained on what you can do when utilizing
a programming language. Python is an incredible path for you to use so as to
acclimate and to do some truly astounding things without getting terrified at
how all the code will look. For certain individuals, a large portion of the
dread of utilizing a programming language is the way that it is difficult to
study and understand all the brackets and other complex issues. In any case,
this isn't an issue with regards to utilizing Python in light of the fact that the
language has been wrapped up to help everybody read and take a look at it
together.
This manual will have all the tools that you have to hit the further
developed parts of Python. Regardless of whether you are seeing this book
since you have a touch of experience utilizing Python and you need to do a
couple of things that are more advanced, or you are beginning as a novice,
you make certain to discover the appropriate responses that you need in the
blink of an eye. So read this manual and discover everything that you have to
know to get some incredible codes while utilizing the Python programming.

Machine learning
What is machine learning?
Machine learning is the act of programming systems to study and learn
from data. For instance a system with artificial intelligence will be able to
determine when an information is true/false, important or spam, etc.
Why machine learning?
Assuming you prefer to write a filter program without machine learning
procedures, you will have to study the problems then write rules and evaluate
them to ensure they are as expected after which you launch the program if
there where error you then have to analyze the problem and repeat the rest
processes and launch again, you observe that every step is done manually
unlike when using artificial intelligent. Also machine learning procedures are
used to solve more complex problems automatically, lastly machine learning
will assist us to learn why machine learning algorithms makes it easy for us
to see what we have learnt.
When would it be a good idea for you to utilize machine learning?
• When you have a problem that requires many not insignificant list of rules
to discover the solution. For this situation, AI procedures can disentangle
your code and improve execution.
• Very perplexing issues for which there is no solution with a customary
approach.
• Non-stable environments: Machine learning programming can adjust to new
information.
Sorts of Systems of Machine Learning
There are various types of machine learning systems. We can partition them
into classes, based upon whether
• They have been trained with humans or not

Supervised
Unsupervised
Semi-administered
Reinforcement Learning
• If they can adapt gradually
• If they work just by contrasting new information to discover data points, or
can identify new patterns in the information,and afterward will manufacture a
model.

What is Supervised Machine Learning?

Most practical machine learning uses supervised learning this type of
machine learning, you train the machine using data which is well "labeled."
Meaning that some data are already tagged with the correct answer,
supervised machine learning is similar to learning which takes place in the
presence of a supervisor.
A supervised learning algorithm learns from labeled training data (where you
have input variables [X] and an output variable [Y] {Y=f(X)}, which helps
you to predict outcomes for unforeseen data. Successfully building, scaling,
and deploying accurate supervised machine learning Data science model
takes time and technical expertise from a team of highly skilled data
scientists. Although, Data scientist will rebuild models to make sure the
insights given remains true until its data changes.
Why Supervised Learning?
Below, are basic reasons for using supervised Learning
Supervised learning helps you to solve various types of real-world
computation problems.
 Supervised learning permits you to collect data or produce a data
output from the previous experience.
Assists you to optimize performance criteria using experience

How Supervised Learning works

Assuming, you want to train a machine to help you predict how long it will
take you to drive home from your workplace. Here, you start by creating a set
of labeled data. This data includes
Time of the day
Weather conditions
Holidays

All these details are your inputs. The output is the time it took you to drive

back home on that day.

As a human you know that if it's raining outside, then the amount of
time spent to drive home will increase. But the machine depends on data and
statistics to co relate outcomes. Let's see how you can develop a supervised
machine learning model of this example which helps the user to determine
the commute time. The first thing you need to create is a training data set.
This training set will contain the total commute time and corresponding
factors like weather, time, holidays etc. Based on this training set, your
machine might see there's a direct relationship between the amount of rain
and time you will take to get home. So, it ascertains that the more it rains, the
longer you will be driving to get back to your house. It might also see the
connection between the time you leave work and the time you'll be on the
road. The closer you're to 6 p.m. the longer time it takes for you to get home.
Your machine may find some of the relationships with your labeled data.
This is the beginning of your Data Model, it starts to reason how rain impacts
the way people drive. It also starts to understand that more people travel
during a particular time of day.

Types of Supervised Machine learning method

Regression: Regression method predicts an output value using training data.
For instance: You can use regression to predict the house price from training
data. The input variables will be locality, size of a house, etc. then a
regression problem has an output variable that has a real value.
Classification:
Classification means to group the output inside a class. If the algorithm
attempts to label input into two distinct classes, it is called binary
classification. Selecting between more than two classes is referred to as
multiclass classification. For instance: Determining whether or not someone
will be a defaulter of the loan.
Supervised machine learning algorithms
Some popular examples of supervised machine learning algorithms are:

Linear regression for regression problems.

Random forest for classification and regression problems.
Support vector machines for classification problems.

Benefit of this model: Outputs always have a probabilistic interpretation, and

the algorithm can be regularized to avoid overfitting.
Disadvantage of this model: Logistic regression may underperform when
there are multiple or non-linear decision boundaries. This method is not
flexible, so it does not capture more complex relationships.

Unsupervised Learning
Unsupervised learning is a machine learning technique, where there no need
to supervise the model. Instead, you will allow the model to work on
independently to predict accurately (here there is absence of output variable).
Unsupervised leaning majorly deals on unlabelled data, unsupervised
learning algorithms permits you to perform more complex processing tasks
than supervised learning. Also unsupervised learning can be more
unpredictable compared with other natural learning deep learning and
reinforcement learning methods.
Why Unsupervised Learning?
Here, are basic reasons for using Unsupervised Learning:
Unsupervised learning finds all kind of unknown patterns in data.
It is simple to get unlabeled data from a computer than labeled
data, which needs manual intervention.
Unsupervised methods help you to find features that can be useful
for categorization.
It occurs in real time, so all the input data to be analyzed and
labeled in the presence of learners.

How Unsupervised Learning works?

Using the case of a baby and her family pet (dog) as an illustration, the baby
knows and identifies this dog. A few weeks later a family friend brings along
a dog and tries to play with the baby. Though the baby has not seen this dog
earlier, but he recognized many features (2 ears, eyes, walking on 4 legs) as
part of her pet dog. She identifies a new animal like a dog. This is naturally
unsupervised learning, where you are not taught but you learn from the data
(in this case data about a dog.) Assuming what happened was supervised
learning; the family friend would have told the baby that it's a dog.
Types of Unsupervised Machine Learning Techniques
Unsupervised learning problems further grouped into clustering and

association problems. Clustering

Clustering is a basic concept when it comes to unsupervised learning. It
mainly deals with finding a structure or pattern in a collection of
uncategorized data. Clustering algorithms will process your data and find
natural clusters(groups) if they exist in the data. You can also change how
many clusters your algorithms should identify. It allows you to adjust the
granularity of these groups.
Association
Association rules permits you to establish associations amongst data objects
inside large databases. This unsupervised technique is about discovering
exciting relationships between variables in large databases. For example,
people that buy a new home most likely to buy new furniture.
Other Examples:
A subgroup of cancer patients grouped by their gene expression
measurements
Groups of shopper based on their browsing and purchasing
histories
Movie group by the rating given by movies viewers

Unsupervised learning algorithms

Some popular examples of unsupervised learning algorithms are:

Apriori algorithm for association rule learning problems. The

Apriori algorithm is used in a transactional database to get
frequent item sets and then generate association rules. It is
commonly used in market basket analysis, where one checks for
combinations of products that frequently co-occur in the database.
In general, we write the association rule for ‘if a person purchases
item X, then he purchases item Y’ as : X -> Y. Example: if a
person purchases milk and sugar, then she is likely to purchase
chocolate powder. This could be written in the form of an
association rule as: {milk,sugar} -> chocolate powder. Association
rules are generated after crossing the threshold for support and
confidence.

k-means for clustering problems. K-means is an iterative algorithm

that groups similar data into clusters. It calculates the centroids of
k clusters and assigns a data point to that cluster having least
distance between its centroid and the data point.
Principal Component Analysis (PCA) is used to produce data easy
to explore and visualize by reducing the number of variables. This
is done by capturing the maximum variance in the data into a new
coordinate system with axes called ‘principal components’.Each
component is a linear combination of the original variables and is
orthogonal to one another. Orthogonality between components
 indicates that the correlation between these components is
zero.The first principal component captures the direction of the
maximum variability in the data. The second principal component
captures the remaining variance in the data but has variables
uncorrelated with the first component. Similarly, all successive
principal components (PC3, PC4 and so on) capture the remaining
variance while being uncorrelated with the previous component.

Semi-Supervised Machine Learning

When you have a large amount of input data (X) and only some of the data is
labeled (Y) you are doing with semi-supervised learning problems. These
problems are in between both supervised and unsupervised learning and this
is where common machine language problems fall into. An example of semi
supervised machine learning is a photo archive where few of the images are
labeled, (e.g. house, plant, animal) and the majority are unlabeled.
Reinforcement learning:
Reinforcement learning is a sort of machine learning algorithm that allows an
agent to decide the best next action based on its current state by learning
behaviors that will maximize a reward.
Reinforcement algorithms usually learn optimal actions through trial and
error. Imagine, for example, a video game in which the player needs to move
to certain places at certain times to earn points. A reinforcement algorithm
playing that game would start by moving randomly but, over time through
trial and error, it would learn where and when it needed to move the in-game
character to maximize its point total.
Batch Learning
In this sort of machine learning, the system can't learn gradually: the system
must get all the required information . That implies it will require numerous
resources and an enormous measure of time, so it's constantly done offline.
Along these lines, to work with this sort of learning, the principal activity is
to train the system, and afterward dispatch it with no learning.
Online Learning
This sort of learning is something contrary to batch learning. I imply that,
here, the framework can learn gradually by furnishing the system with all the
accessible information as instances (individually or groups), and afterward
the framework can learn at once. You can utilize this sort of system for issues
that require the constant progression of data, which needs to adjust rapidly to
any changes. Additionally, you can utilize this sort of system to work with
enormous data indexes, You should know how quick your system can adjust
to any adjustments in the information's "learning rate." If the speed is high,
implies that the system will learn very, rapidly, however it likewise will
overlook old information rapidly.
Instance based learning
This is the least difficult kind of discovering that you ought to learn by heart.
By utilizing this kind of learning in any program, it will execute any
command given to it.
Model-based learning
This kind of learning is based on examples where the system makes
prediction from the available examples.

Insignificant Features

The system might have the option to learn if the training information contains
enough features and information that aren't excessively insignificant. The
main piece of any machine learning project is to grow acceptable features "of
feature engineering".
Feature Engineering
The cycle of feature engineering goes this way:

1. Selection of features: choosing the most valuable features.

2. Extraction of features: consolidating existing features to give more
valuable features.
3. Production of new highlights: making of new highlights, in light of
 information.

System Testing
In the event that you'd prefer to ensure that your model is functioning
admirably and that model can sum up with new cases, you can evaluate new
cases with it by putting the model in the environment and afterward
observing how it will perform. This is a great technique, yet on the off chance
that your model is lacking, the user will grumble. You should isolate your
data into two sets, one set for training and the second one for testing, so you
can train your model utilizing the first and test it utilizing the second. The
speculation error is the rate of blunder by evaluation of your model on the
test set. The value you get will let you know whether your model is
acceptable enough, and in the event that it will work appropriately. In the
event that the error rate is low, the model is acceptable and will perform
appropriately. Conversely, in the event that your rate is high, this implies
your model will perform poorly and not work appropriately. My
recommendation to you is to utilize 80% of the data for training and 20% for
testing purposes, so that it's extremely easy to test or assess a model.
Overfitting the Data
Assuming you visited a home and one of them steals from you, why leaving
the house you will have the thought that every member of that family is a
thief. This is an overgeneralization, and, in machine learning, is designated
"overfitting". This implies that machines do something very similar: they can
perform well when they're working with the training data, yet they can't sum
them up appropriately. For instance, in the accompanying figure you'll locate
a high level of life fulfillment model that overfits the data, yet it functions
admirably with the training data.
When does this happen?
Overfitting happens when the model is extremely complex for the amount of
training data given.
Solutions
To take care of the overfitting issue, you ought to do the accompanying:
- Gather more data for "training data"
- Reduce the commotion level
- Select one with fewer parameters
Underfitting the Data
From its name, underfitting is something contrary to overfitting, and you'll
experience this when the model is exceptionally easy to learn. For instance,
utilizing the case of personal satisfaction, real life is more complex than your
model, so the expectations won't yield the same, even in the training models.
Solution
To fix this issue:
- Select the most powerful model, which has numerous parameters.
- Feed the best highlights into your algorithm. Here, I'm alluding to feature
enginerring.
- Reduce the constraints on your model.

Chapter 2

System classification
For you to appreciate this chapter you must install python, matplotlib and
scikit-learn
Matplotlib python
Matplotlib is a plotting library for the programming language; it provides an
object-oriented API for embedding plots into applications using general-
purpose GUI toolkits like Tkinter, wxPython, Qt, or GTK+. Matplotlib can be
used in Python scripts, the Python and IPython shell, web application servers,
and various graphical user interface toolkits.

Examples of plots in metplotlib

Below are few samples of plots in metplotlib.

Line Plot
Images
Matplotlib can display images

Arbitrary path
Metplotlib has option for arbitrary paths
Scikit-learn is a machine learning library for Python. It features
different algorithms like support random forests, and k-neighbours,
vector machine, and it also supports Python numerical and
scientific libraries like numpy and scipy.
The Mnist
In this part, you'll go further into grouping systems, and work with the Mnist
data set. This is a group of 70,000 pictures of digits handwritten by students
and employees. You'll see that each picture has a label and a digit that
represents it. This task resembles the "hello, world" case of customary
programming. So every learner to AI should begin with this venture to learn
about the grouping calculation. Scikit-Learn has numerous capacities,
including the MNIST. How about we take a look at the code:
>>> from sklearn.data sets import fetch_mldata
>>> mn= fetch_mldata('MNIST original)
>>> mn
{'COL_NAMES': ['label', 'data'],
Description: 'mldata.org data set: mn-original,
data: array([[0, 0, 0,..., 0, 0, 0],
[0, 0, 0,..., 0, 0, 0],
[0, 0, 0,..., 0, 0, 0],
...,
[0, 0, 0,..., 0, 0, 0],
[0, 0, 0,..., 0, 0, 0],
[0, 0, 0,..., 0, 0, 0]], dataType=uint8),
'tar': array([ 0., 0., 0.,..., 9., 9., 9.])} de
. Description is a key that defines the data set.
. The data key here contains an array with only one row for example, and a
column for each feature.
. This target key contains an array with labels.
Let practice with few of the code:
>>> X, y = mn["data"], mn["tar"]
>>> X.shape
(70000, 784)
>>> y.shape
(70000,)
. 7000 here implies that there are 70,000 images, and each picture has more
than
700 features: "784". because, as should be obvious, each picture is 28 x 28
pixels, you can envision that each pixel is one feature.
We should take another instance from the data set. You'll just need to get an
instance’s feature, at that point make it 26 x 26 array, and afterward show
them utilizing the imshow function:
%matplotlib inline
import matplotlib
import matplotlib.pyplot as plt
yourDigit = X[36000]
Your_image = your_image.reshape(26, 26)
plt.imshow(Your_image, cmap = matplotlib.cm.binary,
interpolation="nearest")
plt.axis("off")
plt.show()
As should be obvious in the picture below, it’s like the number five, and we
can give that a label that reveals to us it's five.
 In the following
figure, you can see more mind boggling grouping task from the MNIST data
set.

Additionally, you ought to make a test set and make it before your data is
assessed.

As earlier mentioned the MNIST data set is in two sets, one for training and
the other for testing.
x_tr, x_tes, y_tr, y_te = x [:60000], x[60000:], y[:60000], y[60000:]
How about we play with your training set as follows to make the cross-
validation to be comparative (with no missing of any digit)
Import numpy as np
myData = np.radom.permutaion(50000)
x_tr, y_tr = x_tr[myData], y_tr[myData]

Presently it's an ideal opportunity to make it sufficiently basic, we'll attempt

to simply recognize one digit, for example the number 6. This "6-detector"
will be a case of the binary classifier, to recognize 6 and not 6, so we'll make
the vectors for this task:
Y_tr_6 = (y_tr == 6)// this implies it will be true for 6s, and false for some
other number
Y_tes_6 = (Y_tes == 6)
From that point forward, we can pick a classifier and train it. Start with the
SGD (Stochastic Gradient Descent) classifier. The Scikit-Learn class has the
capacity of taking care of extremely huge data set. In this example, the SGD
will manage occurrences independently, as follows.
from sklearn.linear_model import SGDClassifier
mycl = SGDClassifier (random_state = 42)
mycl.fit(x_tr, y_tr_6)
to utilize it to distinguish the 6
>>>mycl.prdict([any_digit)]
Performance measures
If you want to evaluate the classifier, this will be more difficult than a
regressor, so
Let us explain how to evaluate a classifier.
In this example, we will use cross-validation to evaluate our model.
sklearn.model_selection Import StratifiedKFold
Form sklearn.base import clone
SF = StratifiedKFold (n = 2, run_state = 40)
For train_index, test_index in sf.split (x_tr, y_tr_6):
Cl = no = clone (sagd_claf)
x_tr_fd = x_tr [train_index]
y_tr_fd = (y_tr_6 [train_index])
x_tes_fd = x_tr [test_index]
y_tes_fd = (y_tr_6 [test_index])
Cl.fit (x_tr_fd, y_tr_fd)
y_p = cl.predict (x_tes_fd)
Print (n_correct / len (y_p))
. We use the stratified fold class to do the stratified sampling that generates
folds containing rations for each class. Next, ever iteration in the code will
form a Clone of the classifier to make predictions on the test fold. And
finally, it will calculate the number of correct predictions and their ratio
. We will now use the cross_val_score function to evaluate the SGDC
classifier by K-fold cross validation. The k fold cross validation will split the
training set into 3 folds, and then it will predict and evaluate on each fold.
Sklearn.model_selection import from cross_val_score
Cross_VaL_Score (Sgd, Clf, X_ Y_Tr_6, CV = 3, Scoring = "Accuracy")
You will find the accuracy ratio of "correct predictions" on the folds.
Let's classify each classifier on each image in the not-6
From sklearn.base Import from Base Estimator
Class never 6 Classifier (Base Estimator):
Def fit (self, x, y = None):
Pass
Def Forecast (Self, X):
Return NP.Zeroz ((lane (X), 1), dtype = bool)
Let's check the accuracy of this model with the following code:
>>> never_6_cl = never6Classifier ()
>>> cross_val_score (never _cl, x_tr, y_tr_6, cv = 3, scoring = "accuracy")
Output: array ([“num”, “num”, “num”])
As for the output, you get no less than 90%: only 10% of the images are 6s,
so we
One can always imagine that an image is not a 6. We will be correct about
90% of the time.
Keep in mind that accuracy is not the best performance criterion for
classifiers, if
You are dealing with skewed data sets.

Confusion matrix
The confusion matrix is a better way to evaluate the performance of your
classifier, measuring performance with confusion matrix, is easy just by
calculating the number of times instances of class X are classified as class Y,
for example. To have the number of times of image classifiers of 6s with 2s,
you should look at the 6th row and 2nd column of the confusion matrix.
Let's calculate the confusion matrix using the cross_val_predict () function.
From sklearn.model_selection Imported from cross_val_predict
y_tr_pre = cross_val_predict (sgd_cl, x_tr, y_tr_6, cv = 3)
This function, such as the cross_val_score () function, performs the K fold
cross_validation, it also returns predictions at every fold. It also gives a clean
prediction for each instance in your training set.
We are now ready to obtain the metrix using the following code.
From sKlearn.metrix import confusion_metrics
Confusion_metrics (y_tr_6, y_tr_pred)
You will find an array of 4 values, "numbers".
Each row represents a class in the matrix, and each column represents a
column predicted class.
The first row is the negative: containing "non-6 images". You can learn a lot
from the matrix.
But there is also a good one, which is interesting to work with if you want to
get The accuracy of positive predictions, which is the precision of classifier
using this equation.
Precision = (TP) / (TP + FP)
TP: The number of true positives
FP: Number of false positives
Recall = (TP) / (TP + FN) "Sensitivity": It measures the ratio of positive
instances.
Recall
>>> from sKlearn.matrics import precision _score, recall_score
>>> precision_score (y_tr_6, y_pre)
>>> recall_score (y_tr_6, y_tr_pre)
It is very common to combine precision and recall only one metric, which is
the F1 score.
F1 means both precision and recall. We can count the F1 score with this
equation:
F1 = 2 / ((1 / precision)) ((1) / recall)) = 2 * (precision * recall) / (precision +
Recall) = (TP) / ((TP) + (FN + FP) / 2)
To calculate the F1 score, just use the following function:
>>> from sKlearn.metrics import f1_score
>>> f1_score (y_tr_6, y_pre)
Recall Tradeoff
To get to this point, you should take a look at the SGDC classifier and how it
should look(Makes decisions regarding classification). It calculates the score
based on that decision function, and then it compares the score with the
threshold. if it is more than this score, it will exemplify "positive or
negative".class
For instance, if the decision is in the threshold center, you will find 4 true +
on the right of the threshold, and only one wrong. So the accuracy ratio will
just be
80%.

In SciKit-Learn, you cannot set thresholds directly. You will need access The
decision scores, which uses predictions and by y calling the decision
function, ().
>>> y_sco = sgd_clf.decision_funciton ([any digit])
>>> y_sco
>>> Threshold = 0
>>> y_any_digit_pre = (y_sco> Threshold)
In this code, the SGDClassifier has a threshold, = 0, to return the same
outcome as the predict () function.
>>> Threshold = 20000
>>> y_any_digit_pre = (y_sco> Threshold)
>>> y_ any_digit_pre
This code will prove that, when the threshold increases, the recall decreases.
y_sco = cross_wall_predict (sgd, cl, x_tree, y_tr_6, cv = 3, method =
”decision
Function)
This is the time to calculate all possible precision and remember the threshold
by calling the precision_recall_Curve () function
Import from sKleran.metrics the precision_recall_curve
precisions, recall, threshold = precision_recall_curve (y_tr_6, y_sco)
Let's now plot precision and recall using Metplotib
Def Plot_pre_re (pre_re_thr):
plt.plot (thr, pre [: - 1], “b—“, label = “precision”)
plt.plot (thr, re [: 1], "g-", label = "recall")
plt.xlabel ("Threshold")
plt.legnd (loc = "left")
plt.ylim ([0,1])
Plot_pre_re (pre, re, thr)
plt.show
ROC
ROC is meant for the receiver operating characteristic and it is a tool that
works well with binary Classifieds.
This tool is similar to a recall curve, but does not plot for precision and recall:
It plots the positive rate and negative rates. You will also be able to work
with FPR, which is the ratio of negative samples. You can imagine if it is like
this (1 - negative rate. There is another concept TNR and that is the
specificity. Recall = 1 – specificity).
Let's play with the ROC curve. First, we have to calculate TPR and FPR, just
by calling the roc-curve () function,
from sKlearn.metrics roc_curve
fp, tp, thers = roc_curve (y_tr_6, y_sco)
After that, you can plot FPR and TPR with Metplotlib. According to the
concept of
The following instructions
Def_roc_plot (fp, tp, label = none):
plt.plot (fp, tp, linewidth = 2, label = label)
plt.plot ([0,1)], [0,1], "k--")
plt.axis ([0,1,0,1])
plt.xlabel (‘This is the false rate’)
plt.ylabel (‘this is the true rate’)
roc_plot (FP, TP)
plt.show
Multi-class classification
We use binary classifiers to differentiate between any two classes, but when
you have the task to distinguish between multiple classes you use something
like a random forest classifier or a Bayes classifiers, which can compare
between more than two. But, on the other hand, SVM (support vector
machine) and linear classifiers act like binary classifiers.
If you want to develop a system that categorizes images of digit into 12
classes(0 to 11) You need to train 12 binary classifiers, and create one for
eachclassifier (such as 4-detector, 5-detector, 6-detector and more), and then
you will need to get DS, "decision score" of each classifier for the image.
Then you will choose the highest score classifier. We call this the OvA
strategy: one-versus-all.’’ The other method is to train the binary classifier for
each pair of digits; For instance, one for 5s and 6s and another for 5s and 7s. -
We call this method OvO, "one-versus-one" – to count the number of
classifiers needed depends on the number of classes that use the following
equation: ‘’N= number of classes’’ N * (N-1) / 2. If you want to use this
technique with MNIST 10 * (10-1) / 2,The output will be 45 classifiers,
"binary classifiers".
In Scikit-Learn, you run OvA automatically when you use binary.
classification algorithm.
>>> sgd_cl.fit (x_tr, y_tr)
>>> sgd_cl. predict([any issue])
Additionally, you can call the decision_function () to return scores "10
scores".For a class ”
>>> any_digit_scores = sgd_cl.diction_function ([any_digit])
>>> any_digit_scores
Arrays (["num", "num", "num", "num", "num", "num", "num" "
, ”“ Num ”]])
Training to random forest classifiers
>>> forest.clf.fit (x_tr, y_tr)
>>> Forest.clf.predict ([any digit]]
Array ([num])
As you can see, training with Random Forest Classifier is just two lines of
code and its very simple. Scikit-Learn does not run any OvA or OvO
functions because this type of algorithm - "Random Forest Classifier" - can
perform multiple tasks automatically classes. If you want to take a look at the
list of classifier possibilities, you can call the predict_oroba () function.
>>> forest_cl.predict_proba ([any_digit])
Array ([[0.1, 0, 0, 0.1, 0, 0.8, 0, 0, 0]]))
The classifier is very accurate with its prediction, as you can see in the
output; there is 0.8 at the index number 5.
Let's evaluate the classifier using the cross_val_score() function.
>>> cross_val_score(sgd_cl, x_tr, y_tr, cv = 3, scoring = "accuracy")
Array ([0.84463177, 0.859668, 0.8662669])
You will get 84% more times. When using a random classifier, you will find
in this case, 10% for the accuracy score. Keep in mind that the higher this
value is, the Better.

Error analysis
First of all, when developing a machine learning project:
1. Determine the problem;
2. Collect your data;
3. Work on and explore your data;
4. Clear data
5. Work with many models and choose the best;
6. Connect your models to the solution;
7. Show your solution;
8. Run and test your system.
First, you should work with the confusion metrics and make predictions by
cross-val function. Next, you will call the confusion matrix function:
>>> Y_tr_lover = cross_val_prediciton(sgd_cl, x _tr_scaled, y_tr, cv = 3)
>>> cn_mx = confusion_metrics(y_tr, y_tr_pre)
>>> cn_mx
Array ([5625, 2, 25, 8, 11, 44, 52, 12, 34, 6],
[2, 2415, 41, 22, 8, 45, 10, 10, 9],
[52, 43, 7443, 104, 89, 26, 87, 60, 166, 13]
[47, 46, 141, 5342, 1, 231, 40, 50, 141, 92]
[19, 29, 41, 10, 5366, 9, 56, 37, 86, 189]
[73, 45, 36, 193, 64, 4582, 111, 30, 193, 94],
[29, 34, 44, 2, 42, 85, 5627, 10, 45, 0],
[25, 24, 74, 32, 54, 12, 6, 5787, 15, 236]
[52, 161, 73, 156, 10, 163, 61, 25, 5027, 123]
[50, 24, 32, 81, 170, 38, 5, 433, 80, 4250]]

Plt.matshow (cn_mx, cmap = plt.cm.gray)

Plt.show ()
First, you should divide each value in the matrix by the number of images in
class, and then you will be able to compare the error rate.
rw_sm = cn_mx.sum(axis = 1, keepdims = true)
nm_cn_mx = cn_mx / rw_sum
The next step is to make all the zeros on the diagonal, and it forestall the
errors from occurring.
np.fill_diagonal (nm_cn_mx, 0)
plt.matshow (nm_cn_mx, cmap = plt.cm.gray)
plt.show ()
It is easy to see the errors in the schema above. One thing to keep in mind is
that the rows represent classes and columns represent the predicted values.

Multi-label classifications
In the above examples, there is only one example in each class, let’s say you
want to assign the examples to multiple classes (like facial recognition), for
example
Suppose you want to find more than one face in a single photo. There will be
a label for each face. Let's practice with a simple example.
y_tr_big = (y_tr> = 7)
y_tr_odd = (y_tr% 2 == 1)
Y_multi = np.c [y_tr_big, y_tr_odd]
kng_cl = KNeighborsClassifier()
Kng_cl.fit (x_tr, y_m, ulti)
In these instructions, we have created a y_multi array that has two labels for
each image.And the first contains information on whether the digit is"large"
(8,9,) another checks to see if it's not.
Next, we will make predictions using the following instructions.
>>> kng_cl.predict ([any digit])
Array ([false, true], dataType = bool)
True here means it’s false, meaning it’s not big.

Multi-output classification
At this point, we can cover the final type of classification function, which is
multi-output classification.
It's just a common case of multi-label taxonomy, but each label will have one
multiclass. In other words, it will have more than one value.
Let’s be clarified by this example, using MNIST and adding images some
noise to the image with NumPy functions.
No = rnd.randint (0, 101, (lane (x_tr), 785)))
No = rnd.randint (0, 101, (lane (x_tes), 785))
x_tr_mo = x_tr + no
x_tes_mo = x_tes + no
y_tr_mo = x_tr
y_tes_mo = x_tes
Kng_cl.fit (x_tr_mo, y_tr_mo)
cl_digit = kng_cl.predict (x_tes_mo [any index]])
Plot_digit (cl_digit)
How to train a model
Having gone this far in learning machine learning models and as well
training algorithms which at the first sight was quite challenging, out of this
fear we managed to penetrate regression system, and also cleared the
atmosphere on image classifiers. Although in the previous chapters we
achieved more than that but we didn’t comprehend the core principles of a
model so now we need to progress deeper into the wonder land so that we can
understand how models work and how to use them for purposes.
Getting a deeper understanding of the above details will help you find
the right model and by choosing the best training algorithm. Also, it will help
you with debugging and error analysis. In this chapter, we will deal with
polynomial regression, which is a complex model that works for nonlinear
data sets. In addition, we will work with some regularization techniques that
cuts down training that promote overfitting.

Linear regression
For example, we would take l_S = θ0 + 1 × GDP_per_cap. This is a simple
model
For the linear function of the input feature, select "GPD_per_cap". (θ0 and
θ1) are the dimensions of the model,
In general, you will use a linear model to calculate and predict
the weighted amount of input features and the constant “bias” as written
below
. Y is the value of the predictor.
. N represents the features
. X1 is the value of the feature.
. Θj is the model dimension of J Theta.
Also, we can write an equation in vector form like the following example.

. Θ is the value that lowers the price.

. Y has values from y (1) to y (m).
Let's write some code to practice.
Import numpy as NP
V1_x = 2 * np.random.rand (100, 1)
V2_y = 4 + 3 * V1_x + np.random.randn (100, 1)
After that, we will calculate the value Θ using our equation. It's time to use
this inv () function from our linear algebra module of numpy (np.linalg) to
calculate the inverse of any matrix, and also, the dot () function for
multiplication of our matrix
Value 1 = np.c_ [np.ones ((100, 1)), V1_x]
myTheta = np.linalg.inv (Value1.T.dot (Value 1)). Dot (value 1.T) .dot
(V2_y)
>>> my Theta
Array ([[num], [num]])
This function uses the following equations - y = 4 + 3x + sound “Gaussian” -
To generate our data.
Let's make our predictions now.
>>> V1_new = np.array ([[0], [2]])
>>> V1_new_2 = np.c_ [np.ones ((2,1%), v1_new]]
>>> V2_predicit = V1_new_2.dot (myTheta)
>>> V2_predict
Array ([[4.219424], [9.74422282]])
Now, it’s time to plot the model.
Plt.plot (v1_new, v2_predict, "r-")
Plt.plot (V1_x, V2_y, “b.”)
Plt.axis ([0,2,0,15])
Plt.show ()

Computational complexity
With the general formula, we can calculate the inverse of M^T.M - i.e., a
n * n matrix (n = number of features). The complexity of this inversion is
something like O (n ^ 2.5) to O (n ^ 3.2), depending on the implementation.
Indeed, if you create features like twice, you will create the time the
calculation comes between 2 ^ 2.5 and 2 ^ 3.2.
The good news here is that this equation is a linear equation. Which means it
easily handle a large training set and fit the memory.
After training your model, predictions will not be slow, and the complexity
will be simple, thanks to the linear model. It's time to dig dip and move on
into the procedures of training a linear regression model, which is always
used when large number of features and instances in the memory.
Gradient descent
This algorithm is a common algorithm used for optimization and providing
the best solution for various problems. The idea of this algorithm is to work
with the parameters in an iterative way, to make the cost function very
simple.
The gradient descent algorithm calculates the gradation of error using the
parameter theta, and it works with procedure of descending gradient. If the
gradient is equal to zero, you will reach the minimum.

Also, you should bear in mind that the size of the steps is vital for this
algorithm, because if small ( meaning the rate of learning) is slow, i.e it will
take longer time to cover everything that it needs to.
If the rate of learning is high, it will take less time to cover what’s needed,
and it gives optimal solution.

At times not all cost functions are easy there are some irregular functions that
make getting an optimal solution very hard. This problems are detected when
the local minimum and global minimum looks like the do in the following
figure
If you assign any of the two points on your curve, you will observe that
segment of the line will not join them at the same turn. This cost will look
like a task Bowl, which will happen if the features have many scales like the
image below.
Batch gradient descent
If you want to implement this algorithm, you must first calculate the gradient
of your cost function using theta parameters. If the value of the parameter
Theta has changed; you will need to know the changing rate of your cost
function. We can call this change by a partial derivative
We can calculate the partial derivation using the equations below:
But we will also use the following equations to calculate the partial
derivatives and the gradient vector together.

Let us implement the algorithm.

Lr = 1 # Lr for learning rate
Num it = 1000 # number of iterations
L = 100
myTheta = np.random.randn (2,1)
for it in range (num_it)
gr = 2 / L * Value1.T.dot (Value1.dot (myTheta) - V2_y)
myTheta = myTheta - Lr * gr
>>> My theta
Array ([[num], [num]])
If you try to change the value of the learning rate, you will get different
shapes, e.g.

Stochastic gradient descent

You will encounter a problem when you are using a batch gradient descent: it
requires the use of the whole training set at each step, and that will affect the
performance "speed".
But when using stochastic grad descent, the algorithm will randomly choose
an instance from the training set at any step, and then it will count the value.
The stated procedure makes the stochastic grad descent faster than the batch
gradient descent, knowing that the stochastic gradient descent does not need
the whole set to calculate the value. Another thing to take note of is that
because of its randomness procedure, it will be irregular when compared to
batch algorithm.

Let us implement the algorithm.

Nams = 50
L1, L2 = 5, 50
Def lr_sc (s):
return L1 / (s + L2)
myTheta = np.random.randn (2,1)
for Num in range (Nums):
For l in range (f)
myIndex = np.random.randint(f)
V1_Xi = value 1 [myindex: myindex + 1]
V2_yi = V2_y [myIndex: myIndex + 1]
gr = 2 * V1_xi.T.dot (V1_xi.dot (myTheta) - V2_yi)
Lr = lr_sc (num * f + i)
myTheta = myTheta - Lr * gr
>>> My theta
Array ([[num], [num]])
Mini-batch gradient descent
Because you already know the batch and stochastic algorithms, this kind of
algorithms is very easy to understand and work with. As you know, both
algorithms, calculate the value of a gradients, based on the whole training set
or just one instance. However, the mini-batch calculates its algorithms based
on small and random sets, and performs better than the other two algorithms.
Polynomial regression
We will use this technique, in particular, when dealing with more complex
data. In the case of linear and nonlinear data. After we have add the power of
each feature, we can train the model with new features. This is known as
polynomial regression.
Now, let's write some code.
L = 100
V1 = 6 * np.Random.Rand (L, 1) - 3
V2 = 0.5 * V1 ** 2 + V1 + 2 + np.random.randon (L, 1)
So from the diagram, the straight can never represent the data in the most
efficient way. So we will use the polynomial process to work on this
problem.
>>> from sklearn.preprocession Polynomial import polynomial features
>>> PF = multicellular features (degree = 2, inclusion_seed = false)
>>> V1_P = PFFit_Transform (V1)
>>> V1 [0]
Array ([number])
>>> V1_P [0]
Now, let's do this work properly with our data, and change the straight line.
>>> ln_r = Linear Argument ()
>>> ln_r.fif (V1_P, V2)
>>> ln_r.intercep_, ln_r.coef
Learning curves
Suppose you are dealing with polynomial regression, and you need it to fit
the data better than that of linear one. In the image below, you will find 300-
degree model. We can also compare the final result with other types of the
regression: "normal linear".
In the figure above, you can see the overfitting of data when you’re using the
polynomial. On the other hand, with linear one, you can see that the data
apparently been underfitted.

Regularized Linear models

We have worked in the first and second chapters on how to reduce overfitting
by regularizing the model a little, as an example, if you’d to like to regularize
a Polynomial model. In this case, to fix the problem, you should decrease the
number of degrees.
Ridge regression
Ridge regression is another version of linear regression, but, then regularizing
it and adding weight to the cost function, this makes it fit the data, and makes
the weight of the model as smooth as possible. Here is the cost function of
the ridge regression:

As an example of ridge regression, just observe the following figures.

Lasso regression
“Lasso” regression stands for “least absolute contraction and selection
operator”regression. This is another type of regularized version of the linear
regression.
It sounds like a ridge regression, but there is a small difference in the
equation, as in the following figures
The cost function of the Lasso Regression Operations:
As you can see in the figure below, Lasso regression uses smaller values than
the ridge.
Chapter 4

Combinations of different models

Tree classifiers
The next image illustrates the definition of a common target of collecting
functions that are simply for merging different classifiers into one-class
including a more generalization performance than each individual classifier
alone.
For example, suppose you have collected predictions from many experts.
Ensemble methods will allow us to merge these predictions by lots of experts
to get a prediction that is stronger than the predictions of each individual
expert. As you can see later in this chapter, there are a lot of different
methods for creating a combination of classifiers. In this part, we will
introduce one basic understanding of how ensemble works and why they are
common to achieve a good generalization performance.
In this chapter, we will work with the most popular ensemble method that
uses the majority voting principle. A lot of voting simply means we choose
the label that has been predicted by most classifiers (more than 50% votes
was achieved). For example, the word vote here is just binary class settings
only. However, generating a majority voting principle to multi-class setting is
not difficult
The principle is multi-class settings, called majority voting. After that, we
The class will select the label that received the most votes. The following
figure
Explains the concept of majority and majority voting for the 10 alliance
Classified where each unique symbol (triangle, square and circle) a
Unique class label:
Using the training set, we start by training m different classifiers (CC1,,…
8M).
Depending on the method, the ensemble can be made from several
classifications algorithms; For example, decision trees, support vector
machines, logistics regression classifier, and so much more. In fact, you can
use the same base classification algorithm fitting different subsets of the
training set. An instance of this method will be a random forest algorithm,
which combines many decisions ensemble ways to use majority voting.
To predict a class label by simple majority or plurality voting, we combine
these predicted the class labels of each individual classifier C j and select the
class label yˆ that got the most votes:
y m ˆ = ode {CC 1 2 () x x ,, ()), Cm () x}
For example, in a binary classification function where class 1 1 = - and class
2 1 = +, we can write a majority vote prediction.
Furthermore to explain why ensemble methods will work better than
individual classifications alone, let's apply the simple concepts of
combinatory. For the following example, we assume that all n base classifiers
for binary classification task have the same error rate, ε . Furthermore, we
assume that the classifiers are independent and does not relate to the error
rate. As you can see, we can only explain the error statistics of an ensembe of
the base classifiers as probability.

Mass function of a binomial distribution:

Here, n. k is the binomial coefficient: n choose k. You can see how simple it
is to calculate the probability that the prediction of the ensemble is incorrect.
Now, let's work out a more solid example of 11 base classifiers (n = 11) with
an error rate of 0.25 (ε = 0.25):
You may notice that the error rate of the ensemble (0.034) is lower than the
error rate of each individual classifier (0.25) if all assumptions are met. Note
that in this simplified picture, a 50-50 split by an even number of classifier n
is considered as an error, where this is the correct half of the time. To
compare such an idealistic nsemble classifer to a base classifer over a range
of dfferent base error rates, let’s implement the probability mass function in
python:
>>> Import math
>>> def ensemble_error (n_classifier, error):
... q_start = math.ceil(n_classifier / 2.0)
... probability = [comb (n_class, q) *
... error ** Q *
... (1-error) ** (n_classifier - q)
... for q in series (q_start, l_classifier + 2)]
... return sum (probability)
>>> ensemble_error(n_classifier = 11, error = 0.25)
0.034327507019042969
Let's write some code to calculate the rate for visualization of different errors
the relationship between ensemble and base errors in a line graph:
>>> import numpy as np
>>> error_Range = no.arrange (0.0, 1.01, 0.01)
>>> en_error = [en_er (n_classifier = 11, er = er)
... in er in er_rang]
>>> Import matplotlib.pyplot as plt
>>> plt.plot (er_range, en_error,
... label = ' ensemble error’,
... linewidth = 2)
>>> plt.plot (er_range, er_range,
... ls = '--', label = 'B_ er',
... linewidth = 2)
>>> plt.xlabel ('B_ er')
>>> plt.label ('B / En_er')
>>> plt.legend (loc = 'top left')
>>> plt.grid ()
>>> plt.show ()
As we can see in the following plot, the error probability of an ensemble is
usually better than the error of an individual base classifier as long as the
base classifiers perform better than random predictions (ε <0.5). You should
consider that the y-axis shows the base error as well as the ensemble error
(continuous line):
Implementing simple majority classifier
As we saw in the introduction to merge learning in the last section, we will
work using a warm-up training and then create a simple classifier for a
majority voting in Python programming. As you can see, the next algorithm
will work on multiclass settings by plurality voting; you will use the term
majority voting for simplicity as common in most articles.
In the following program, we will develop and also join different
classification programs associated with individual weights for confidence.
Our goal is to create a strong meta-classifier that balances out the individual
classifiers’ in a particular dataset. In more precise mathematics onditions, we
can write the weighted majority vote.
To translate the idea of a weighted majority vote into Python code, we can
Use NumPy friendly Argmax and bincount functions:
>>> import numpy as np
>>> np.Argemax (np.bincount ([0, 0, 1]),
... weight = [0.2, 0.2, 0.6]))
.
Here, pij is the predicted probability of jth classifier for class label i. To
Continue with our previous example, let's assume we have a binary
Classification problem with the connection of class labels i ∈ {} 0,1 and an
ensemble of 3 classifiers Cj (J ∈ {} 1,2,3). Let us assume that Classifier Cj
returns
The following class membership probabilities for a specific sample x:
C1 () x → [ ] 0.9,0.1, C2 () x → [ ] 0.8,0.2, C3 () x → [ ] 0.4,0.6
To implement a weighted majority vote based on class probabilities, we can
also use NumPy using numpy.average and np.argmax:
>>> ex = NP array ([[0.9, 0.1],
... [0.8, 0.2],
... [0.4, 0.6]])
>>> P = np.average (ex, axis = 0, weight = [0.2, 0.2, 0.6])
>>> p
Array ([0.58, 0.42])
>>> np.argemax (P)
0
Putting everything together, let's now apply a majority vote classifier to it
Python:
from Sklearn.base import classifiermixin
from sklearn.pre_processing import from label_En
from sklearn.ext import six
from sklearn.ba import clone
from sklearn.pipline import _name_estimators
Import numpi as np
Import operator
Class MVClassifier (BaseEstimator,
ClassifirMixin):
"" "A majority vote ensemble classifier
Parameters
----------
cl: array-like, shape = [n_classifiers]
Different classifiers for the ensemble votes: str, {'cl_label', 'prob'}
Default: 'cl_label'
The 'cl_label' prediction is based on the argmax of the class labels. Elif
'Probe', The arg of the total probs is used to predict the class label
(recommended for calibrated classifier).
W: arr-like, s = [n_cl]
Optional, default: None
If a list of `int or float values is provided, the classifiers are weighed by the "
"
def __init __ (s, cl,
v = 'cl_label', w = none):
s.cl = cl
s.name_cl = {key: value for
Key, value in
_name_estimators (cl)}
s.v = V
s.w = w
def fit_CL (s, X, y):
"" "Fit_cl.
Parameters
----------
X: (array-like, sparse matrix},
s = [n_samples, n_features]
Matrix of training samples.
y: arr_like, sh = [n_samples]
Vector of target class labels.
Returns
-------
s: object
"" "
# use LabelEncoder to ensure class labels start
# With 0, which is important for np.argmax
# call in s.predict
s.l_ = LabelEncoder()
s.l_fit (y)
s.cl_ = self.lablenc_.classes_
s.cl_ = []
for cl in s.cl:
fit_cl = clone(cl).fit (X,
s.la_.transform(y)
s.cl_.append(fit_cl)
return s
I added many comments to the code to better understand the individual parts.
However, before implementing the rest of the methods, let’s take a quick
break and discuss some of the code that may be confusing at first. We use
parent classes Base Estimator and ClassifierMixin to get some support
functionality for free, including get_params and set_params to set and return
the classifier’s parameters as well as the score method for calculating the
prediction accuracy, respectively. Also, note that we imported six to create
the MajorityVoteclassifier compatible with Python 2.7.
Next, we will add the prediction method of predicting a class label by a
majority vote based on the class labels, if we start a new
majorityvoteclassifier object with votes = 'classlabel'. Alternatively, we'll be
able to get started the ensemble classifier with vote = 'probability' to predict a
class label based on class membership probability. In addition, we will add a
prediction_proba method to return the average probabilities, which is useful
to calculating the Receiver Operator Characteristic Area under the curve
(ROC AUC).
Def Pre (s, X):
""" Pre-class labels for X.
Parameters
----------
X:{arr- like, spar mat},
Sh = [n_ samples, n_ features]
Matrix of training samples.
Returns
----------
ma_v: arr-like, sh = [n_samples]
Predicted class labels.
"""
If se.v == 'probability':
ma_v = np.argmax (spredict_prob(X),
Axis = 1)
else: #ˈcl_labelˈ v
predictions = np.asarr ([cl.predict (X))
For cl in
s.cl _]). T.
ma_v = np.ap_al_ax (
Lambda X:
np.argmax (np.bincount (x, weight = s.w)),
Axis = 1,
arr = forecasts)
ma_v = s.l_.inverse_transform(ma_v)
return ma_v
def predict_proba(self, X):
"""prediction for X.
parameters
----------
X: {arr-like, sp mat},
sh = [n_samples, n_features]
Training vectors, where n_sample is the number of samples and n_samples
Number of features.
Returns
----------
av_prob: like array,
sh = [n_samples, n_classes]
Average weight probability for each class per sample.
"""
Probes = np.asarr ([cl predict_probe(x)
for cl in s.cl_])
av_prob = np.av (probes,
axis = 0, weight = s.w)
return av_prob
def get_ps (self, deep= true):
"""Get categorized parameter names for Gridsearch"""
If not deep:
return super (MVC,
self) .gat_ps (deep=False)
else:
ou = s.n_cl.copy ()
For n, step in\
Six.iteritems (s.n_cl):
for k, value in six. iteritems (
step.get_ps (deep=true)):
ou ['% s __% s'% (n, k)] = value
return ou
Combining different algorithms for classification with majority vote
Now, it’s almost time to put in the MVC we used earlier in action. You
should first create a dataset on which you can test. We are already familiar
with the techniques for loading datasets from CSV files, We will take a
shortcut and load the Iris dataset from scikit-learn's dataset module.
Also, we will choose only two features, the width of the sepals and the length
of the petal make the classification task a bit difficult. Although our
multiclass classifier or MVC normalizes to multiclass problems, we will
classify only flower samples from two classes, Ir-versicolor and Ir-Virginica
to calculate ROC AUC. The code is as follows:
>>> Import sklearn as sk
>>> Import sklearn.cross_validation as CV
>>> Ir = datasets.load_ir()
>>> X, y = ir.data [50:, [1, 2]], ir.target [50:]
>>> le = label Encoder()
>>> y = le.fit_transform(y)
Next we divide the Iris samples into 50 percent training and 50 percent test
data:
>>> x_train, x_test, y_train, y_test =\
... train_test_split (X, y,
... test_size = 0.5,
... random_state = 1)
Utilizing the training dataset, we now train three different classifiers a logistic
regression classifier, a decision tree classifier, and a K-nearest neighbors
classifier and study their various performance through a 10 cross-validation
Before we merge them into ensemble one:
Import the following
sklearn.cross_validation
sklearn.linear_model
sklearn.tree
sklearn.pipline
Pipeline
numpy as np
>>> clf1 = logistic regression(penalty = 'l2' ',
... C = 0.001,
... random_state =0)
>>> clf2 = DTCL (max_depth = 1,
... criterion = 'entropy',
... random_state = 0)
>>> cl = KNC (n_nb = 1,
... p = 2,
... met = 'minsk')
>>> pipe1 = pipeline ([['sc', StandardScaler()],
... ['clf', clf1]])
>>> pipe3 = pipeline ([['SC', standardscaler()],
... ['clf', clf3]])
>>> clf_labels = ['Logistics Regression', 'Decision Tree', 'KNN']
>>> Print ('10-fold cross validation:\n ')
>>> For clf, label in zip ([pipe1, clf2, pipe3], clf_labels):
... SC = CrossVSc (estimator = clf,
>>> X = X_train,
>>> y = y_train,
>>> cv = 10,
>>> scoring='roc_auc')
>>> Print ("ROC AUC:% 0.2f (+ / -%0.2f) [% s]"
...% (scores.mean(), scores.std(), label)
As shown in the following snippet, the output we receive shows that the
predictive performance of an individual classifier is almost same:
10-fold cross validation:
ROC AUC: 0.92 (+/- 0.20) [Logistic regression]
ROC AUC: 0.92 (+/- 0.15) [Decision Tree]
ROC AUC: 0.93 (+/- 0.10) [KNN]
You may be wondering why we have trained the logistics regression and K-
nearest neighbors classifier as part of a pipeline. The reason here is, as we
said, Logistic regression and K-near neighbors algorithms (using Euclidean
distance metric) is not scale-invariant unlike with decision trees. Though, the
Ir benefits are all measured on an equal scale; It is a good habit to work with
standard features.
Now, let’s proceed to the more exciting part and connect the individual
classifiers for majority rule voting in our M_V_C:
>>> mv_cl = M_V_C (
... cl = [Pipe 1, clf2, pipe 3])
>>> cl_labels + = ['Majority voting']
>>> all_cl = [pipe1, clf2, pipe3, mv_clf]
>>> For cl, label in zip (all_clf, clf_labels):
... sc = cross_val_score(est=cl,
... X = X_train,
... y = y_train,
... cv = 10,
... scoring = 'roc_auc')
...% (scores.mian (), scores.study (), label)
R_AUC: 0.92 (+/- 0.20) [Logistic regression]
R_AUC: 0.92 (+/- 0.15) [D_T]
R_AUC: 0.93 (+/- 0.10) [KNN]
R_AUC: 0.97 (+/- 0.10) [majority vote]
In addition, the output of the MajorityVotingClassifier has substantially
improved over the classifiers in 10-fold cross-validation evaluation.
Classifier
In this section, you are calculating the R_C curve from the test set to check if
MV_classifier generalizes well to unseen data. We must remember that
testing set will not be used for model selection; The only goal is a report an
estimation of the accuracy of the classifier system. Let's take a look at
imports matrix.
import roc_curve from sklearn metrics import auc
cls = ['black', 'orange', 'blue', 'green']
ls = [':', '-', '-.', '-']
For CL, label, cl, 1\
... in zip (all_cl,, cl_labels, cls, ls):
... y_pred = clf.fit (X_train,
... y train) .product_proba (X_test) [:, 1]
... fpr, tpr, thresholds = rc_curve(y_t = y_tes,
... y_sc = y_pr)
... rc_auc = ac (x = fpr, y = tpr)
... plt.plot (fpr, tpr,
... color = clr,
... linestyle = ls,
... la = '% s (ac =% 0.2 f)'% (la, rc_a))
>>> plt.lg (lc = 'lower right')
>>> plt.plot ([0, 1], [0, 1],
... linestyle = '-_',
... color = 'gray',
... linewidth = 2)
>>> plt.xlim ([- 0.1, 1.1])
>>> plt.ylim ([- 0.1, 1.1])
>>> plt.grid ()
>>> plt.xlb ('False positive rate')
>>> plt.ylb ('True Positive Rate')
>>> plt.show ()
As we can see in the resulting ROC, the ensemble classifier also performs
well
On the test set (ROC AUC = 0.95), while K-near neighbors classifier seems
to be over-fitting the training (training ROC AUC = 0.93, test ROC AUC =
0.86):
You select only two features for classification functions. It will be interesting
to show what the decision field of ensemble classifier looks like. However it
is not necessary to standardize the training features before the model to fit
because our logistics regression and K-nearest neighbors pipelines will take
care of this automatically, you will set up training set so that the decision
regions of the decision tree will be on the same scale for visual purposes.
Let's dive in:
>>> sc = SS ()
X_tra_std = sc.fit_transform(X_train)
From itertools imported product
x_mi = X_tra_std [:, 0] .m () - 1
x_ma = X_tra_std [:, 0] .ma () + 1
y_mi = X_tra_std [:, 1] .mi() - 1
y_ma = X_tra_std [:, 1] .ma () + 1
xx, yy = np.meshgrid(np.arange (x_min, x_max, 0.1),
... np arange (y_mi, y_ma, 0.1)
f, axarr = plt.subplots (nrrows = 2, ncols = 2,
Sharex = 'col',
sharey = 'row',
figze = (7, 5)
for ix, cl, tt in zip (product ([0, 1], [0, 1])),
all_cl, cl_lb):
... cl.fit (X_tra_std, y_tra)
... z = cl.f.predict (np.c_ [xx.ravel (), yy.ravel ()])
... z = Z.res (xx.shape)
... character [IDX [0], IDX [1]]. Contu (_xxx, _ ye, z, alf = 0.3)
... character [idx [0], idx [1]]. Scatter (X_tra_std [y_tra == 0, 0],
... X_tra_std [y_tra == 0, 1],
... C = 'Blue',
... score = '^',
... s = 50)
... character [idx [0], idx [1]]. Skatt (X_tra_std [y_tra == 1, 0],
... X_tra_std [y_tra == 1, 1],
... c = 'red',
... marker = 'O',
... s = 50)
... character [idx [0], idx [1]]. set_title (tt)
>>> plt.text (-3.5, -4.5,
... z = 'slay wide [standard]',
... yes = 'center', w = = 'center', ftsize = 12)
>>> plt.text (-10.5, 4.5,
... z = 'p_length [standard]',
... yes = 'center', va = 'center',
... f_size = 13, rotation = 90)
>>> plt.show ()
Interestingly, but as expected, the areas of decision of the class included
The decision of the individual classified class seems to be a hybrid of
regions. At
At first glance, the majority vote seems like a boundary decision
K-nearest neighbor classifier boundary. However, we can see that it is
Like the stump of the decision tree, the orthogonal for the y axis for the
sapphire width-1
Before you learn how to tune individual classifier parameters for
configuration
Classification, let's call the get_ps method to find the essential idea of how
we can
Access the individual parameters inside the grid search object budget:
>>> mv_clf.get_params ()
decision 'decantentriclassifier': decantentriclassifier (class_weight = nothing,
criterion = 'entropy', maximum_depth = 1,
Max_tails = nothing, max_leaf_nodes = nothing, minute_samples_
Leaf = 1,
Minute_samples_split = 2, minute_weat_fraction_leaf = 0.0,
random_state = 0, splitter = 'best'),
decisiontreeclassifier_class_weight': none,
decisiontreeclassifier__'s class': 'entropy',
[...]
decisiontreeclassifier_random_state': 0,
decisiontreeclassifier__splitter: 'best',
'Pipeline-1': pipeline (steps = [('sc', StandardScaler (copy= true, with_
mean = True, with_std = True)), ('clf', Logistic Regression(C = 0.001,
Class_weight=None,
dual = False, fit_incept = true,
intercept_scaling = 1, max_ iter = 100, multi_class = 'ovr',
penalty = 'l2', random_state = 0, solver = 'libliner',
toll = 0.0001,
verbose = 0),
'Pipeline-1__ clf'_Cˈ: 0.001,
'Pipeline-1__ clf'_class_weightˈ: None,
'Pipeline-1__ clf'_dualˈ:False,
[...]
ˈpipeline-1_sc_with_stdˈ: True
'pipeline-2': Pipeline(steps=[('sc', StandardScaler(copy=True, with_
mean=True, with_std=True)), ('clf', KNeighborsClassifier(algorithm='au
to', leaf_size=30, metric='minkowski',
metric_params=None, n_neighbors=1, p=2,
w='uniform'))]),
'p-2__cl”: KNC(algorithm='auto', leaf_
size=30, met='miski',
met_ps=None, n_neighbors=1, p=2,
w='uniform'),
'p-2__cl__algorithm': 'auto',
[...]
'p-2__sc__with_std': T}
Depending on the values returned by the get_ps method, you now know how
to access the individual classifier's attributes. Let’s work with the inverse
regularization parameter C of the logistic regression classifier and the
decision tree depth via a grid search for demonstration purposes. Let’s take a
look at:
>>> from sklearn.grid_search import GdSearchCV
>>> params = {'dtreecl__max_depth': [0.1, .02],
... 'p-1__clf__C': [0.001, 0.1, 100.0]}
>>> gd = GdSearchCV(estimator=mv_cl,
... param_grid=params,
... cv=10,
... scoring='roc_auc')
>>> gd.fit(X_tra, y_tra)
After the grid search has completed, we can print the different hyper
parameter value combinations and the average R_C AC scores computed
through 10-fold cross-validation. The code is as follows:
>>> for params, mean_sc, scores in grid.grid_sc_:... print("%0.3f+/-%0.2f
%r"
... % (mean_sc, sc.std() / 2, params))
0.967+/-0.05 {'p-1__cl__C': 0.001, 'dtreeclassifier__ma_depth': 1}
0.967+/-0.05 {'p-1__cl__C': 0.1, 'dtreeclassifier__ma_depth': 1}
1.000+/-0.00 {'p-1__cl__C': 100.0, 'dtreeclassifier__ma_depth': 1}
0.967+/-0.05 {'p-1__cl__C': 0.001, 'dtreeclassifier__ma_depth': 2}
0.967+/-0.05 {'p-1__cl__C': 0.1, 'dtreeclassifier__ma_depth': 2}
1.000+/-0.00 {'p-1__cl__C': 100.0, 'dtreeclassifier__ma_depth': 2}
>>> print('Best parameters: %s' % gd.best_ps_)
Best parameters: {'p1__cl__C': 100.0,
'dtreeclassifier__ma_depth': 1}
>>> print('Accuracy: %.2f' % gd.best_sc_)
Accuracy: 1.00