FAKE NEWS DETECTION USING MULTI

LAYER PERCEPTRON ALGORITHM

TEAM MEMBERS TEAM GUIDE

BADRINATH M H DR. SUNDAR C
DARWIN PANDIA RAJ V (HOD/CSE)
AHMED SHALIH S
HARI LINGESHWARAN K
 OBJECTIVE

• The main objective is to classify the fake news using deep learning
algorithms with improved accuracy rate

• And also predict the fake news posted in social network

 INTRODUCTION

• Fake news detection refers to the process of identifying false or misleading information that is spread through
traditional and online media channels.

• With the rise of social media and the ease of access to information, it's become increasingly important to be
able to detect fake news and distinguish it from accurate and trustworthy sources.

• There are several techniques used to detect fake news, including fact-checking, examining the source of the
information, looking for evidence to support claims, and using tools such as machine learning algorithms.

• However, it's important to note that no single approach is foolproof and a combination of methods is often
needed to determine the accuracy of a news story.
 ABSTRACT

• Information quality in social media is an increasingly important issue, but web-scale data hinders

experts’ ability to assess and correct much of the inaccurate content, or “fake news,” present in these

platforms.

• Automated detection of fake news is a hard task to accomplish as it requires the model to understand

nuances in natural language.

• Develops a method for automating fake news detection on datasets by learning to predict accuracy

assessments
 EXISTING SYSTEM
• Random Forest Algorithm : which combines the output of multiple decision trees
to reach a single result

• Navies Bayes Algorithm : Calculates the probability of an event

• LSTM model : To provide global features and local word embedding features of
news dataset
 DISADVANTAGES
• Accuracy is less

• Need large number of datasets to train the data

• Provide high number of false positive rate

• Only done supervised classification

 PROPOSED SYSTEM
• "Fake News" is a term used to represent fabricated news or
propaganda comprising misinformation communicated
through traditional media channels

• Can implement text mining algorithm to extract the key

terms based on natural language processing

• And also include classification algorithm such as deep

learning algorithm named as Multi-layer perceptron
ADVANTAGES

• Reduce the false positive rate

• Analyze all types of features

• Improve the accuracy rate

• Time complexity can be reduced

SYSTEM ARCHITECTURE
SYSTEM REQUIREMENTS

• HARDWARE CONFIGURATION

• Processor : Dual core processor 2.6.0 GHZ

• RAM : 1GB

• Hard disk : 160 GB

• Compact Disk : 650 Mb

• Keyboard : Standard keyboard

• Monitor : 15 inch color monitor

SYSTEM REQUIREMENTS

• SOFTWARE CONFIGURATION

• Operating system : Windows OS

• Front End :PYTHON

• Back End : MYSQL

• Application : Web application

MODULES

• TRAIN THE DOCUMENTS

• TEXT MINING

• DOCUMENT TERM MATRIX CONSTRUCTION

• CLASSIFICATION

• FAKE NEWS DETECTION

 MODULE 1
TRAIN THE DOCUMENTS

• Internet contains vast collections of high-quality information, but it often

provides more than what is needed.

• Text summarization is an application of information retrieval that condenses

input text while preserving its meaning and information content.

• Query-specific document summarization using similarity measures has been

extensively researched.

• The module allows users to upload standard text files and collect large news
datasets, enabling efficient selection of data for specific information needs.
 MODULE 2
TEXT MINING

• The text documents in .TXT format are collected as the initial step.

• Document pre-processing is performed, which involves removing

redundancies, inconsistencies, and separating words.

• Tokenization is applied to the document, where the string is divided

into individual words or tokens.

• Stop words, such as common words like "a," "an," "but," "and," "of,"
and "the," are removed.
 MODULE 3
DOCUMENT TERM MATRIX CONSTRUCTION

• The module includes the ability to calculate term frequency (TF) and inverse
document frequency (IDF).

• TF-IDF, which stands for term frequency–inverse document frequency, is a numerical

statistic used to measure the importance of a word in a document within a collection
or corpus.

• TF-IDF takes into account both the frequency of a word in a document and its
frequency in the overall corpus.

• Words that appear frequently in a specific document but infrequently in the entire
corpus are considered more significant.
 MODULE 4
CLASSIFICATION

• The module allows users to input news datasets or Twitter datasets for analysis.

• The implemented algorithm in this module is the multi-layer perceptron (MLP),

which is a type of feedforward artificial neural network.

• MLP consists of multiple layers, including input, hidden, and output layers,
connected as a directed graph. It uses backpropagation for training the network.

• MLP is commonly used for tasks such as prediction, classification, pattern

recognition, and function approximation.
 MODULE 5
FAKE NEWS DETECTION

• The classification of news items into fake or real has gained significant
attention from researchers worldwide.

• Numerous studies have examined the impact of falsified and fabricated news
on individuals and their reactions to such content.

• Falsified news refers to any textual or non-textual content that is intentionally

misleading and created to deceive readers into believing something false.

• The proposed system aims to improve the accuracy of fake news detection by
predicting and classifying fake news data.
SCREENSHOTS
IDE
Prediction Matrix
Web Application Link
Web Application
Detecting Fake News
Prediction Output
 FUTURE ENHANCEMENT

• In future, we can extend the framework to implement various deep learning

algorithms to improve the accuracy and reduce the complexity in classification.
 LITERATURE SURVEY
S.NO TITLE AUTHOR FINDINGS ADVANTAGES DISADVANTAGES
1 Evaluating Deep Learning Apurva The fake news Performed Only support the
Approaches for Covid19 Wani detection task is thorough trained datasets
Fake News Detection formulated as a text experiments on
classification transformer-based
problem. We solely models and
rely on the content sequential models
of the news and
ignore other
important features
 LITERATURE SURVEY
S.NO TITLE AUTHOR FINDINGS ADVANTAGES DISADVANTAGE
S
2 Fake news detection using Bahad, P. The recent growth in to predict fake There is no
bi-directional LSTM- Saxena, the field of machine news article using automated
recurrent neural network 2019 learning also came up deep learning approach to predict
with the theories and models the fake news
algorithms to detect
fake data
 LITERATURE SURVEY
S.NO TITLE AUTHOR FINDINGS ADVANTAGES DISADVANTAGES
3 A sensitive stylistic approach de Oliveira, In this letter, we propose Presented a stylistic- Not implemented in
to identify fake news on social Nicollas, a computational stylistic computational analysis, real time
networking 2020 analysis based on based on natural environments
natural language language processing
processing, efficiently
applying machine
learning algorithms to
detect fake news in texts
extracted from social
media
 LITERATURE SURVEY
S.NO TITLE AUTHOR FINDINGS ADVANTAGES DISADVANTAGES
4 MVAE: Multimodal Dhruv The model consists Trained by jointly Extend MVAE using
Variational Autoencoder Khattar of three main learning the tweet propagation
for Fake News Detection components, an encoder, decoder data and user
encoder, a decoder and the fake news characteristics
and a fake news detector
detector module.
 LITERATURE SURVEY
S.NO TITLE AUTHOR FINDINGS ADVANTAGES DISADVANTAGES
5 Fake news detection using Kumar, Collect 1356 news Compares multiple Does not support
deep learning models: A novel Sachin, 2020 instances from various state-of-the-art newly updated
approach users via Twitter and approaches datasets
media sources such as
PolitiFact and create
several datasets for the
real and the fake news
stories
 LITERATURE SURVEY
S.NO TITLE AUTHOR FINDINGS ADVANTAGES DISADVANTAGES
6 The Role of User Profiles Kai Shu, Study the problem Aim to answer To further
for Fake News Detection Xinyi Zhou of understanding questions regarding understand their
and exploiting user nature and extent of utilities for fake
profiles on social the correlation news detection
media for fake between user
news detection profiles on social
media and fake
news
 LITERATURE SURVEY
S.NO TITLE AUTHOR FINDINGS ADVANTAGES DISADVANTAGES
7 A Survey on Recent Advances Merryton, This paper also showcases classifications used There is no security in
in Machine Learning 2020 a survey on different to identify fake news fake news detection
Techniques for Fake News researches performed in
Detection fake news detection using
traditional machine
learning methods and
Deep Neural Networks
 LITERATURE SURVEY
S.NO TITLE AUTHOR FINDINGS ADVANTAGES DISADVANTAGES
8 Fake news detection regarding Nikiforos The rapid development Described an Computational
the Hong Kong events from of network services has innovative and well- process in low
Tweets led to the exponential defined method for
growth of online detecting fake news in
information and the
increasing number of social media
social media users
 LITERATURE SURVEY
S.NO TITLE AUTHOR FINDINGS ADVANTAGES DISADVANTAGE
S
9 A Sensitive Stylistic Nicollas R. Presented a stylistic- In the process of Does not support
Approach to Identify Fake de Oliveira computational assessing the large datasets
News on Social analysis, based on quality of the
Networking natural language detection of the
processing, efficiently methodologies,
applying unsupervised an accuracy of
learning algorithms, 86%
such as one-class SVM
 LITERATURE SURVEY
S.NO TITLE AUTHOR FINDINGS ADVANTAGES DISADVANTAGES
10 User Preference-aware Yingtong In this paper, we In this paper, we Accuracy is less in
Fake News Detection Dou study the novel argues that user fake news detection
problem of endogenous news
exploiting user consumption
preference for fake preference plays a
news detection vital role in the fake
news detection
problem
REFERENCES
• de Oliveira, Nicollas R., Dianne SV Medeiros, and Diogo MF Mattos. "A sensitive stylistic approach to
identify fake news on social networking." IEEE Signal Processing Letters 27 (2020): 1250-1254.

• Nikiforos, Maria Nefeli, et al. "Fake news detection regarding the Hong Kong events from tweets." IFIP
International Conference on Artificial Intelligence Applications and Innovations. Springer, Cham, 2020.

• Merryton, Adline Rajasenah, and Gethsiyal Augasta. "A survey on recent advances in machine learning
techniques for fake news detection." Test Eng. Manag 83 (2020): 11572-11582.

• Kumar, Sachin, et al. "Fake news detection using deep learning models: A novel approach." Transactions on
Emerging Telecommunications Technologies 31.2 (2020): e3767.

• Bahad, Pritika, Preeti Saxena, and Raj Kamal. "Fake news detection using bi-directional LSTM-recurrent
neural network." Procedia Computer Science 165 (2019): 74-82