CHITTAGONG UNIVERSITY OF ENGINEERING & TECHNOLOGY (CUET)

DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING

CHITTAGONG – 4349
___________________________________________________________________________
(Project Proposal)

Application for the Approval of B.Sc. Engineering Thesis / Project

(Computer Science & Engineering)
Date: 23.09. 2019

1. Name of Student : Abu Bakkar Siddique

Roll No. : 1504110 Session: 2018-2019

2. Present Address : Room No: A-309, Bangabandhu Hall, Chittagong

University of Engineering & Technology, Raozan,
Chittagong

3. Name of Supervisor : Dr. Mahfuzul Hoq Chowdhury

Designation : Assistant Professor
Department of Computer Science & Engineering (CSE)
Chittagong University of Engineering & Technology

4. Name of the Department : Computer Science & Engineering(CSE)

Program : B. Sc. Engineering

5. Date of First Enrollment

in the program : February 25, 2016

6. Tentative Title : Developing a Mental Healthcare Assisting system for

University Students of Bangladesh

1
7. Introduction
According to World Health Organization moreover 450 million people suffer from mental
health problems. This proves that mental disorders are the causes of poor health condition
and disability worldwide [3]. Again, depressive disorders are the fourth leading cause of
global burden and expected to be second leading by 2020 [2]. Depression is a very common
mental illness that negatively affects how a person feels, react and it is very common
nowadays. It causes loss of interest in activities and sadness. It affects a person’s
concentration in work, studies, and daily life. One out of 15 adults suffer from depression
which shows how serious the illness is [5]. Anxiety disorder is another psychological issue
that affects nearly 30 people of adults at some point in their life [5]. It involves anxiety and
fear. Depression, anxiety, stress can make a person hurt themselves badly and have suicidal
thoughts

In Bangladesh, over 6 million people experience depressive disorders and almost 7 million
people have anxiety disorders. It is estimated that more than 10,000 people are dying by
suicide in the country. Among students aged 15-22 in Bangladesh, 4% of boys and 6% of
girls consider attempting suicide. Stigma, discrimination and misconceptions about mental
disorders have a severe impact on health seeking behavior in Bangladesh. Many people
suffer from it and still don’t know it as they don’t know the symptoms.

University students are one of the susceptible groups who are experiencing mental health
problems. Different studies conducted among undergraduate student in different part of
South Asia revealed that they are suffering from psychological morbidity. Even in
Bangladesh undergraduate students suffer from many psychological and mental problems.
The reason behind these problems are many. As Bangladesh is a poverty oriented country
students face many problems such as tuition fee, low family income problem etc. Again
undergrad students often fall victim to ragging or bullying which causes mental trauma.
Some students might have present/past psychological disorder that is getting worse day by
day. Some might have broken family or relationship problems. As a result many undergrad
students are lacking in their academic performance. Students often don’t share these
problems with others, not even with their friends and families out of embarrassment and
social stigma.

So in our paper we propose a system that would diagnose psychological disorders such as
depression and anxiety using machine learning algorithms. First we will construct a
questionnaire that contains a student’s certain information such as family income, present/past
psychological health background, relationship status, academic result, political involvement
etc. to prepare our dataset. Then we will use some machine learning algorithm such as
Convolutional Neural Networks, Support vector Machines, K Nearest Neighbors etc. to find
the optimal algorithm for our precise prediction of mental disorder .It will help to find the
students having mental issues that needs counselling by analyzing the prepared dataset. We will
not only diagnose the problems but will try to give mental health care to the students who are
suffering from mental disorders. For that purpose, we will try to develop a mobile application
that will give mental healthcare to the students in need.

2
8. Background Study
Many research has been done to detect mental disorder and treatment .The author [1] Roland
H.C. Yap ,David M. Clarke described in their paper that using Constraint Logic programming
(CLP) to construct a system that could diagnose mental disorder. However it is a very straight
forward process. The author [] Rozita Masri, Hazar Mat Jani used three artificial intelligence
reasoning techniques , rule based reasoning, fuzzy logic and fuzzy genetic algorithm for the
diagnosis. The work is limited to rule based reasoning and the questionnaires are very lengthy
with many options.
Many research included image processing to study the patients psychological state more
accurately. The author in [7] David De Vault , Ron Artstein , et al. introduced a prototype of a
virtual interviewer that could measure anxiety, depression, post traumatic disorder by analyzing
the patients gesture. The limitation is that it does not provide any mental healthcare or treatment.
Other studies include disorder detection from social media data like twitter, Facebook etc. For
example, author [6] Guangyao Shen, Jia Jia proposed a depressive dictionary model that can
find out depressed people on twitter. They analyzed twitter dataset which differentiated between
depressed from non-depressed people. The paper only finds depressed people but don not take
any actions for the welfare of depressed persons. In another paper, the author [8] Janith
Weerasinghe, Kediel Morales, et al. described that patients with mental sickness can be found
out by language pattern. They also suggested model that can detect mental sickness in twitter.
They used word unigrams, word clusters and feature analysis to detect important feature. Their
main aim was to use language pattern to detect mental illness. This paper is also focused on
finding depression rather than offering any mental assessment.
However every paper that we went through was based on the behavior of a particular region.
Those may show different result if we were to use it in the perspective of our country. Again
most of them emphasized to find mental disorder but very few could provide mental healthcare
to the victims of mental disorder.

9. Objective with Specific Aims and Possible Outcomes

The main objective of this work is to develop a depression and anxiety analysis system model
and later provide medical healthcare to the patients. The key objective and the possible
outcomes of this work may mention in the following:

3
 9.1 Motivation
Most of the university students of Bangladesh suffers from mental disorder for various
reasons such as family income, academic result, relationship problems, political
involvement, drug addiction, ragging/bullying etc. These problems are true for CUET
also. When undergraduate students join here they fall victim to ragging, drug addiction,
relationship problems, and political involvement very fast. So to make them interested to
academic activities, sports and other activities we need a system that can identify students
having depression, anxiety etc. psychological disorder and provide them with necessary
mental support and mental healthcare.
9.2 Objective
 To design and develop a depression and anxiety analysis system that will diagnose
depression and anxiety disorder with high accuracy.
 To evaluate the system model using dataset prepared from survey questionnaires.
 To design and develop a mobile app to provide mental health care to students.

9.3 Possible Outcomes

 We may detect students suffering from depression.
 After a good counselling, they may give proper concentration to their studies and
daily life.
 CUET administration might adopt this system to provide mental healthcare to its
students and workers.
10. Outline of Methodology/Experimental design
Mental disorder prediction module:

1. Anxiety
User Questionnaire Dataset
2. Depression

Machine learning
algorithms (K Nearest
Result after Final Dataset
Neighbour/CNN/Random
analysis (After cleaning data)
forest/SVM/Linear
Regression)

Figure-1: System model for disorder prediction

4
The proposed model is described here:
1. The whole procedure starts with questionnaires asked from users. The questionnaires
will have following information:
 Family income/support
 Present/past psychological health background
 Academic result
 Political involvement
 Any previous disciplinary punishment
 Relationship problem
 Drug addiction
 Victim to ragging/bullying etc.
2. Then user’s data will be collected in a excel sheet for further use.
3. Filtering will be done by removing incomplete data and finalize it for data analysis.
4. Then machine learning algorithms such as Linear Regression, K Nearest Neighbor,
Support Vector Machine, Linear Discriminant Analysis, Convolutional neural Network
etc. The reason for using different algorithms is that different algorithms work better in
different cases. Some work better with less data, whereas some are effective with huge
data depending on the algorithms specialty. We want our module to give us the best
accuracy possible to determine mental disorder.
5. Then by comparative analysis we will find the optimal algoritm.

After finding the students with mental disorder we need to design our mobile application to
provide mental healthcare to the students. In fig.2 the working model of the application is
shown

5
Healthcare application working model:

Student analyzed as depressed

Analyze the questionnaire to find

the possible reason of depression

Family Academic Drug Bullying or Others

Income Result Addiction Ragging

Scholarshi Extra care Rehabilitatio Contact with Other

p/Part time in studies n the provost recommend
job recommenda ations
tion

Figure-2: Working model of the application

The proposed model is described below:

1. When a student is detected as depressed or having mental disorder the app will
analyze the student’s questionnaire again and ask the student the specific reason of
depression.
2. Based on the students answer the application will provide mental healthcare such
as
 If a student says he is suffering from mental disorder for family income
problem the app will contact the university to offer the student scholarship,
part-time job or sanction a bank loan.
 For academic result issues the app will contact the students’ advisor or
course teacher for extra care in studies.

6
  If the answer is present/past psychological disorder the app will prove the
student a hotline to get connected with an expert psychiatrist or counselling
expert
 For drug addiction problem the app will contact the student’s hall
provost/legal guardian and recommend a good rehabilitation center.
 For ragging/bullying the app will contact the hall provost first, then contact
the law enforcement as it is a crime and provide the student with expert
mental counselor recommendation.
In this way the app will try to provide mental healthcare to students
suffering from mental disorder as many way possible.

11. Resources Required Completing the Work:

The necessary tools to implement this project can be divided in to two Categories-Hardware
and Software as illustrated below:
 Hardware Requirements

 Personal Computer

 Software Tools

 For Implementing Algorithm :

 Python, Anaconda and other packages.
 Android or PC based software development environment

 Cost Estimate:
The costs that will occur to implement our proposed system are given bellow:

7
 Cost of Materials:
A general laptop Tk. 45,000
Paper Tk. 500
Total Tk. 45,500
Typing, Drafting and Binding:
Internet Browsing & Typing Tk. 2,000
Drafting Tk. 500
Binding Tk. 500
Total Tk. 3,000
Grand Total Tk. 48,500

12. References

[1] R. Yap and D. M. Clarke, “An expert system for psychiatric diagnosis using the dsm-iii-r,
dsm-iv and icd-10 classifications.", in Proceedings of the AMIA Annual Fall Symposium,
American Medical Informatics Association, p. 229,1996
[2] World Health Report (September 21, 2019) [Online].
Available: “https://www.who.int/whr/2001/media centre/press_release/en/”,2001
[3] W. H. Organization et al., “Investing in mental health", 2003.
[4] F. Deeba and R. Begum, “Development of an anxiety scale for Bangladesh population",
Bangladesh Psychological Studies, vol. 14, pp. 39-54, 2004.
[5] American psychiatric association (September 21, 2019) [Online].
Available: “https://www.psychiatry.org/” ,2019
[6] R. Y. Masri and H. M. Jani, “Employing artificial intelligence techniques in mental health
diagnostic expert system", International Conference on Computer & Information Science
(ICCIS), IEEE, vol. 1, pp. 495-499, 2014
[7] G. Shen, J. Jia, L. Nie, F. Feng, C. Zhang, T. Hu, T.-S. Chua, and W. Zhu,
“Depression detection via harvesting social media: A multimodal dictionary learning
solution", in IJCAI, pp. 3838-3844, 2017
[8] J. Weerasinghe, K. Morales, and R. Greenstadt, “Analyzing machine learning models that
predict mental illnesses from social media text", International Conference on Computer &
Information Science (ICCIS),IEEE, vol. 3, pp. 57-69, 2019

8
13. CSE Undergraduate Student (CUGS) Committee reference:

Meeting No: Resolution No: Date:

14. Number of Under-Graduate Student(s) working with the Supervisor at Present:

Signature of the student

________________________
Signature of the Supervisor

Signature of the Head of the Department