Volume 7, Issue 11, November – 2022 International Journal of Innovative Science and Research Technology

ISSN No:-2456-2165

Assistive Mobile Application for Software Engineers

in Sri Lanka to Support Depression ‘Emoods’
Shashika Lokuliyana Anuradha Jayakody Chiranthi Ranasinghe Raveen Jayasena
Department of Computer Department of Computer Coventry University UK Department of Computer
Systems Engineering, Systems Engineering, Psychiatry Department Systems Engineering,
Faculty of Computing, SLIIT Faculty of Computing, SLIIT Colombo South Teaching Faculty of Computing, SLIIT
Malabe, Sri Lanka Malabe, Sri Lanka Hospital Malabe, Sri Lanka

Jeynika Tharmaratnam Hansani Rajapaksha Javindu Kumarasiri

Department of Computer Department of Computer Department of Computer Systems
Systems Engineering Systems Engineering Engineering
Faculty of Computing, SLIIT Faculty of Computing, SLIIT Faculty of Computing, SLIIT
Malabe, Sri Lanka Malabe, Sri Lanka Malabe, Sri Lanka

Abstract:- This paper presents an Assistive mobile I. INTRODUCTION

application in Sri Lanka to support depression. The
framework uses face recognition technologies and Depression is a mood disorder that affects millions of
algorithms to identify depression prediction via machine people around the world. Lately, most with the Covid-19
learning for the users. The most effective means of crisis arising around the globe, people being stuck to one
improving the quality of Depression and mental illnesses place, persons have become so much exposed to severe
at work is becoming increasingly widespread in the tech depression conditions, withnegative repercussions. Thus, it is
industry. Software developers, according to the vital to have a good supporting system around such people to
International Journal of Social Sciences, have a far higher help them get over thecondition. Other than a close circle of
risk of depression, burnout, anxiety, and stress than their friends and family, a professional mental health doctor or a
colleagueswho execute mechanical activities. Employees' therapist could get you on the right path by providing an
mental health, as well as the company's total productivity, accurate diagnosis and the best treatment options. There are
is threatened by declining mental health. Researchers certain phases of life that an individual has to get through and
from Stuttgart's Institute of Software Technologies most commonly, this chronic illness does contain different
discovered that developers who are emotionally types of depression, listed as bipolar disorder, major
exhausted or depressed generate lower-quality code and depressive disorder, postpartum depression, and post-
are more concerned about missing deadlines. The traumatic stress disorder. An estimated 3.8 percent of the
objective of this research is to determine the prevalence of world's population suffers from depression, with 5.0 percent of
depression among Sri Lankan software engineers. It is adults and 5.7 percent of persons over 60 years old being
critical not to deal with depression on one's own. They affected. In 2020, there will be 280 million depressed persons
require a system of loving individuals, such as family around the globe. In Sri Lanka, thereare an estimated 802,321
members, friends, coworkers, and neighbors, who enable cases overall, or 4.1 percent of the population. By age,
them to be themselves. Building and maintaining a strong prevalence rates change, reaching a peak in later adulthood
support system of people who can provide (above 7.5 percent among females aged 55-74 years, and
encouragement, help to keep moving and involved in above 5.5 percent among males) [1].
meaningful activity, and help them challenge their
negative thinking is a critical part of an assistive mobile Although depression is curable, fewer than a quarter of
app. This app provides features such as patient attention, people obtain proper treatment of mental conditions because
patient awareness, treatment for patient depression of treatment access challenges such as time and transportation
levels, monitoring patient progress through time series restrictions, lengthy waiting lists, and a shortage of qualified
analysis, collecting patient information via chatbot, and specialists to provide high-quality care. The use of
monitoring the improvement of doctor-patient technology might significantly reduce access issues; several
relationships. studies have already shown that Internet-based methods are
as possible and efficacious as in-person therapy. Due to the
Keyword:- Machine Learning, Depression, Facial effectiveness of these remote methods, there is a lot of interest
Expression Analysis, Treatment, Chatbot in using mobile applications as a substitute foundation for the
delivery of healthcare.

IJISRT22NOV492 www.ijisrt.com 831

Volume 7, Issue 11, November – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
Even though there are IT solutions for depression across functional and useful[5].
the world, Sri Lanka does not have any. Because of the
pressure to deliver rapidly and the assumption that difficult According to another research, Smartphones and other
issues will haveimmediate answers, the majority of software pervasive and personal sensing technologies have made it
engineers are at greater risk and are frequently pushed to possible to continuously gather data in an inconspicuous way,
lower their typical standards. It may be detrimental to certain allowing for this to be done in real-time. To anticipate user
creators' sense of purpose. In order to help Sri Lankan contextual information such as location, mood, physical
software engineers overcome this, we decided to create an activity, and so on using machine learning techniques,
aiding application. In a nutshell, these Mental health continuous sensor data have been used in machine learning
applications provide users with the freedom to use these methods. Recently, there has been an increasing interest in
resources whenever they need to and as frequently as they utilizing pervasive sensing technology for applications in
like, not having to wait for a mental health expert to be mental health care. This would allow for the automated and
present. continuous monitoring of various mental illnesses, such as
anxiety, stress, depression, and so on. The purpose of this
Finding out how often depression is among Sri Lankan studyis to provide an overview of current research efforts in
software developers is the purpose of this research. A crucial the field of mental health monitoring systems (MHMS),
component of an assistive mobile app is creating and which use sensor data and machine learning. Our primary
maintaining a solid network of support that may offer concentration was on research works pertaining to mental
motivation, aid in keeping users active and engaged in diseases and conditions such as stress, anxiety, depression,
worthwhile activities, and support in challenging their and bipolar disorder, among others. We offer the main stages
maladaptive cognitions. Our major goal is to create a mobile of MHMS as well as a categorization taxonomy that we
application that includes all of these and that genuinely suggest using as a reference for reviewing related
supports individuals in their struggle with depression. The publications. In addition, the difficulties encountered by
Chatbot element in this assistive mobile app acts as a sentient researchers working in the sector aswell as the potential for
touch point and offers a special opportunity to improve the future are highlighted[6].
accessibility and user-friendliness. A chatbot is a type of
virtual assistant that can communicate and engage in Students who are in need of mental health care confront
graphical, textual, and spoken dialogue with human users. a number of obstacles, including the expense of treatment, its
Theuser-friendly chatbot is automated and put into place for location, its availability, and the stigma associated with
their comfort. Additionally, it strives to improve user seekingtreatment. Studies have shown that computer-assisted
experience byaccelerating user responses and responding to therapy and one conversational chatbot that gives cognitive
all queries. behavioral therapy (CBT) offer an alternative treatment that
is less time- consuming and more cost-effective for the
II. LITERATURE REVIEW treatment of anxietyand depression. Applying an integrated
approach has been associated with similarly effective
According to “Development of a Mobile Phone App to posttreatment improvement as cognitive behavioral therapy
SupportSelf-Monitoring of Emotional Well-Being: A Mental (CBT), which is widely considered to be one of the most
Health Digital Innovation” [2], The state of one's emotional successful treatment modalities. Integrative psychological
health is critical to both one's mental health and overall well- artificial intelligence (AI)provides a solution that is scalable
being. Using conventional methods, however, it is in response to the growing need for help that is not only
challenging to keep track of the day-to-day changes in an economical but also convenient, long-lasting, and safe [7].
individual's emotional state over a long period of time[3].It is
difficult to provide support for mental health when only The most recent developments in mobile
around one in two personswho have mental health problems communications and technology have the door to brand new
seek professional assistance.The technology that is included opportunities for mobilehealth (mHealth)[8]. These gadgets,
in mobile phones provides a way that may be maintained to of which there are more than 1 billion cellphones and 100
improve one's ability to self- manage their emotional well- million tablets throughout theglobe, have the potential to be
being. The purpose of this study is to discuss the creation of a useful tools in the administration of health care. Because of the
tool for use on mobile phones that is intended to monitor more than 50 million fatalities that were predicted to be
changes in an emotional state in arealistic daily setting and in caused by diseases or health issues in 2008, any assistance that
real-time. This evidence-based mobile phone application can be provided for medical treatmentis not only appreciated
monitors the user's emotional and mental health and well- but also required. Depending on how common some of these
being, and it gives connections to the websites of illnesses are certain ones take on a greater level of
organizations that focus on mental health as well as other significance.
resources. Self-report psychological questionnaires,
experience sampling methodology (ESM),[4], and automated Detecting depression in its earliest stages is critical and
behavioral data collection are the three methods that are used may possibly save a patient's life. Nonlinear analysis of EEG
togather information using this app. The results of a survey signals is examined in this work to distinguish between
and focus group discussions with 11 persons (ages ranging depressed patients and healthy controls. This research
from 16to 52 years old; 4 males and 7 females) who used the included 45 depressive individuals who were unmedicated
app for a period of thirty days revealed that it was both and 45 healthy volunteers. From the EEG data, we were able

IJISRT22NOV492 www.ijisrt.com 832

Volume 7, Issue 11, November – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
to extract the power of four EEG bands and four nonlinear and deep learning techniques. The results of the test,
characteristics, such as detrended fluctuation analysis (DFA), including whether or not the individual is depressed and,
Higuchi fractal, correlation dimension, and Lyapunov if so, the level of depression, would be revealed next,
exponent. k-nearest neighbor, linear discriminant analysis, followed by the indicators gained from the machine
and logistic regression are employed as classifiers to learning algorithms derived via those activities, such as the
distinguish between the two groups. In addition to the duration of a patient would take to complete the activity,
correlation dimension and LR classifier, the highest his/her facial expressions, and the scored amount of the
classification accuracy of 83.3% may be achieved using given activity. In order to construct a questionnaire to
nonlinear features. Classifiers use a combination of all measure an individual's stress levels, it is necessary to use
nonlinear characteristics to boost their performance even a mobile application development tool.
more. The LR classifier and all nonlinear characteristics work
together to obtain a classification accuracy of 90%.[9] The B. Create a system for analyzing and curing: An automated
genetic algorithm is used in every test toidentify the most therapy suggestion system is being trained using user input
significant characteristics. Comparing and contrasting the data gleaned from surveys and interviews with patients
suggested strategy with existing techniques, it is shown that about their demographics and mental health. To teach
incorporating nonlinear characteristics improves machine learning models to make sequential judgments,
performance. An EEG nonlinear analysis may be used to tell one might use a technique called reinforcement learning.
apart depressed individuals from healthy ones, as shown in The agent learns to function in a harsh and unpredictable
thisresearch. Psychiatrists might use this data in addition to environment. The topic of reinforcement learning uses a
their current methods of identifying depression in their gaming environment to teach computers new skills. The
patients. algorithm uses trial and error to determine the best course
of action. Artificial intelligence may be rewarded or
Both anxiety and depression are different mental health penalized for specific actions to encourage it to perform the
conditions that are presently diagnosed based on the patient's intended actions.
self-reporting of their symptoms. Subclinical anxious and
depressed persons may now be diagnosed using a novel C. Creating a system to avoid aloneness: Using chatbots,
diagnostic approach that systematically examines for which are powered by AI, is a scalable option that provides
cognitivebias abnormalities. A total of 125 individuals were an engaging way to engage customers in behavioral health
categorizedinto four groups depending on the severity of their therapy. Natural language processing (NLP) is used to
anxiety anddepression symptoms. A wide range of cognitive recognize text while users converse with the chatbot, and
and emotional biases was discovered and measured using a NLP is also used to comprehend user conversations as the
complete battery of behavioral tests. In this work, advanced mobile app is being developed. Then, you may anticipate
machine-learning technologies were used to examine the a language analysis and user language identification. It is
data. In order to predictgroup membership, these techniques also expected that you would employ a number of
use specific characteristics that distinguish anxiety and strategies to achieve this goal. A chatbot is an AI-powered
depression. When comparing people with high symptoms of conversational virtual assistant designed to take over
depression, anxiety, or both to those without, the prediction mundane, repetitive tasks such as answering common
model showed a sensitivity of 71.44% and a specificity of questions or resolving common problems. In order to
70.78%, respectively, for the two groups (specificity). Two- forecast the user's response, artificial intelligence (AI) use
group models with highdepression/anxiety had a prediction machine learning algorithms to decipher the user's natural
accuracy of 68.07 percent and 74.18 percent, respectively. language input. Machine learning algorithms to decipher
The analysis also revealed which particular behavioral the user's naturallanguage input.
measures contributed to the prediction and pointed to critical
cognitive pathways in anxiety vs depression. The study D. Creating a database and analyzing the patient feedback:
findings In light of these findings, newdiagnostic devices, and User input, both good and negative, as well as suggestions
personalized treatment plans may be developed. for improvement, is essential for reaching this goal. We
can categorize their comments as good, negative, or
III. RESEARCH OBJECTIVES constructive depending on how they choose to respond to
us. All users have the option of responding to comments
The main objective of this research is to provide a or receiving star ratings for their responses. All comments
unique solution for the current depression among software are being recorded in a database for future analysis. The
engineers in Sri Lanka and to develop a mobile application to purpose of this survey is to help the segment's creators
really support them in depression. learn how to better serve those experiencing depression
and use that knowledge in the future. The feedback
A. A method for identifying the depression of the patient: analysis feature will make use of sentiment analysis to
Python, PyCharm, and a variety of databases are used to decipher the text of a Google form, Natural Language
determine the degree of a person's depression. Face Processing (NLP), and the flutter framework.
recognition technology can be used to identify and monitor
patients. While patients are participating in the specified
activity, a face recognition tool will be used to ascertain
their emotional state using conventional neural networks

IJISRT22NOV492 www.ijisrt.com 833

Volume 7, Issue 11, November – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
IV. METHODOLOGY After that, this system will begin providing chatbot
functionality. Using that chatbot feature, the patient may talk
with the system in order to gain an understanding of the mobile
application, and the functional working details. The greatest
part of this chatbot is that it offers live chat about his or her
depression, and it will recognize the patient's depression. In
theend, the system will give a feedback mechanism, which
will be utilized in the upcoming innovations that we want to
accomplish. It analyzes both the good and the negative
feedback that has been given. Overall, this mobile application
uses image processing, natural language, developing
languages, and many other tools are used for the development
of the mobile application.

A. Identification of depression level.

To identify the depression level system will generate
basic questionnaires. Face recognition finds and tracks
Fig 1: Overall System Diagram patients. During the procedure of patients partaking in the
High-level architecture specified activity, anticipate employing a face recognition
tool to determine their facial expressions. Indicators gained
This section provides an overview of the solution's from machine learning algorithms derived from those
overall design as well as its evolution. An investigation is activities, such as the duration a patient would take to
conducted into the conceptual foundation of the system, complete the activity, his/her facial expressions, and the
including its capabilities. After that, the visual specification, scored amount of the given activity, would give the results of
technical creation, and overall system design of the suggested the test; whether the individual is suffering from depression
solution are presented to the audience. and, if so, the level of depression. A mobile app development
tool is needed to create a stress questionnaire. Here is the
The "Assistant mobile application for software survey.
engineers in Sri Lanka to assist depression" system research
focuses on how theframework makes use of facial recognition Over the last two weeks, how often have you been
technologies and algorithms for identifying depression bothered byany of the following problems?
prediction for the users using machine learning. There are 1. Little interest or pleasure in doing things
primarily four different partsthat make up the solution.  Not at all
 Several days
A. Identification of depression level.  More than half of the days
B. Treatment Suggest system.  Nearly every day
C. Developing chatbot. 2. Feeling down, depressed, or hopeless
D. Analyzing user feedback component.
 Not at all
 Several days
System Design: This Assistive mobile application for
 More than half the days
software engineers in Sri Lanka to support depression has
developed for 4 components as follows;  Nearly every day
3. Trouble falling or staying asleep, or sleeping too much
The primary use of this is to serve as a depression-level  Not at all
system for software developers in Sri Lanka. Regarding this  Several days
particular scenario, we are delivering a mobile application  More than half the days
that has several different functions. The system's ability to  Nearly every day
recognize the level of deterioration is the very first feature of 4. Feeling tired or having little energy
the system. In light of this, the following depression-level-  Not at all
related questions have been provided for your consideration.  Several days
During that period of time, the technology that recognizes  More than half the days
facial expressions will be operational. The image processing  Nearly every day
algorithm is used to determine the amount of depression a 5. Poor appetite or overeating
person is experiencing. By utilizing these two different  Not at all
approaches, the system is able to provide some treatments  Several days
that the user may carry out on herself or herself. A therapy  More than half the days
suggestions system is what you are looking at right now. Not  Nearly every day
only will the treatment suggestion system continue to 6. Feeling bad about yourself - or that you are a failure or
functionfor that, but it will also determine the location of the havelet yourself or your family down
treatment facility that is geographically closest to the patient.
 Not at all
That is going to serve as a proposal.
 Several days

IJISRT22NOV492 www.ijisrt.com 834

Volume 7, Issue 11, November – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
 More than half the days reward, starting with random trials andrising to complicated
 Nearly every day tactics and superhuman abilities. Reinforcement learning
7. Trouble concentrating on things, such as reading the uses search and trials to indicate computer inventiveness.
newspaper or watching television Artificial intelligence can learn from thousands of
 Not at all simultaneous gameplays provided a strong enough computer
 Several days infrastructure is used.
 More than half the days
 Nearly every day In the proposed system, the model will be trained with
8. Moving or speaking so slowly that other people could the sample data to analyze the questionnaire and from the
havenoticed analysis, the model will suggest treatment for the patients.
 Not at all
 Several days
 More than half the days
 Nearly every day
9. Thoughts that you would be better off dead, or of hurting
yourself
 Not at all
 Several days
 More than half the days
 Nearly every day
10. If you've had any days with issues above, how difficult
have these problems made it for you at work, home, Fig 2: Methods for treatment suggest the system
school, or with other people?
C. Developing chatbot.
 Not difficult at all (0)
A scalable option that offers an interactive way to
 Somewhat difficult (1)
engage consumers in behavioral health therapies powered by
 Very difficult. (2) artificial intelligence is chatbots. Developing the mobile
 Extremely difficult. (3)Total - /30 application is using natural language processing (NLP) to
identify text when users chat with the chatbot and Natural
Depression Severity: 0-4 none, 5-9 mild, 10-14 Moderate, 15- Language Processing to understand the dialogues of users.
19 moderately severe, 20-27 severe. Not at all – 0 Then, expect to identify language and analysis the language
Several days -1 that language using users. And also expected to do that by
More than half the days -2Nearly every day -3 using several methods. Chatbots are conversational virtual
assistants that employ artificial intelligence to automate
For that purpose, we obtain the inputs such as facial interactions with humans. Artificial intelligence (AI) is used
recognition, questionnaire duration time, and the score of the to predict replies to the users, which uses machine learning
questionnaire, from the user to our system. we anticipate techniques to interpret natural language.
recognizing whether the user is suffering from depression or not
and if yes, which level of depression he or she is at. Moreover, D. Analyzing user feedback component.
the facial recognition system of the application is also In order to do this, it is needed to collect positive,
connected to the back end of the diagram, which is further negative, andconstructive feedback from users and analyze
connected to the database that has all the gathered details of the userfeedback. Based on their responses we can segment
thepatients. those feedbacks as positive, negative, and constructive. All
the users can be given feedback by posting a comment or
The Details that the user enters into the system go to the rating their feedback as stars. Feedback is stored in the
Backend part and are saved in the database. database and we are looking to analyze each and every
feedback. The Objectiveof collecting feedback is to apply for
B. Treatment Suggest system. future improvements in the depression level and identify the
First, patients will be provided questionnaires to gather segment. It is going to be used sentimental analysis to read
user input data to assess their depression level depending on the words in the google form,Natural Language Processing
age andgender. Then, the reinforcement learning model will (NLP), and flutter framework to develop the feedback
evaluate the data to train to auto-recommend a therapy. analysis component.
Reinforcement learning teaches machine learning models to
make judgments. The agent learns to attain a goal in an IV. RESULT AND DISCUSSION
uncertain environment. Reinforcement learning puts AI in a
game-like setting. The computer tries things till it works. AI This section provides the details of the results of the
is rewarded or punished for activities to encourage the system. Basically, this system is depending on the
machine to do what the programmer wants. It aims to creativeness of the mobile application and also on the back-
maximize return. The designer sets the reward policy (game end development of themobile application. If we discuss the
rules) but gives the model no guidance on how to win. The UI of the system the first page provides the details of the name
model must learn how to fulfill the task to maximize the with other details ‘that have headings called “How is my

IJISRT22NOV492 www.ijisrt.com 835

Volume 7, Issue 11, November – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
depression right now?”,” Chat with Ms.Deppresia”, and also calming music, journaling, connecting with mental health
the “Sign Up” icon. resources, seeking assistance, meditation, and mood tracking
are employed. It would be helpful to analyze the functionality
of mobile apps for treating depression and any ethical issues
they may raise given their rapid rise. We propose that
marketplace developers regulate depression applications to
eliminate ethical concerns including missing, weak, or
inconsistent privacy regulations, such as sharing data with
third parties, safeguarding kid data, and protecting vulnerable
user groups. The app analysis offered new knowledge on how
to prevent theharm caused by hazardous material, unrestricted
access (with accompanying privacy problems), and less
accurate screening procedures.

REFERENCES

[1]. P. A. A. H. K. G. A. T. J. J. C. A. D. J. H. P. &. A. A. J.
Arean, " The Use and Effectiveness of Mobile Apps for
Fig 3: Logging page and Chatbo Depression," December 2016, pp. 1–6
[2]. H.A. A., D. B. ,. E. S. Nikki Rickard, "JMIR Mental
The Chatbot is provided as “Chat with Ms. Deppresia” health," JMIR publications, 23 November 2016. pp.23
the patient can chat with the system, and they can ask about [3]. M. EnriqueGarcia-Ceja, "Pervasive and Mobile
their system and about their level of depression and many Computing," December 2018, pp. 18-27
other things about their depreciation. This is basically like [4]. A. J. ,. B. G. L. L. Russell Fulmer, "Using Psychological
“WhatsApp chat” but advanced with a face recognition Artificial Intelligence to Relieve Symptoms of
system. Depression andAnxiety: Randomized Controlled Trial,"
JMIR Production, 1312 2018, pp. 64-68
[5]. M. A. Gamage, R. M. Arachchi, S. Naotunna, T.
Rubasinghe, C. Silva and S. Siriwardana, "Academic
Depression Detection Using Behavioral Aspects for Sri
Lankan University Students," 2021 3rd International
Conference on Advancements in Computing (ICAC),
2021, pp. 335-340, DOI:
10.1109/ICAC54203.2021.9671214.
[6]. I. d. l. T.-D. M. L.-C. Borja Martínez-Pérez, "Mobile
Health Applications for the Most Prevalent Conditions
by the World Health Organization: Review and
Analysis," JMIR Publications, 14 6 2013.
[7]. B. F. M. G. R.-L. Thalia Richter, "Using machine
learning- based analysis for behavioral differentiation
between anxiety and depression,” Scientific Reports,
October 2020.
Fig 4: Depression level and feedback system [8]. R. M. A. S. N. R. C. S. Madhuransi Agrahere Gamage,
"Academic Depression Detection Using Behavioral
These UIs are providing the feedback and depression Aspects for Sri Lankan University Students," IEEE
level measuring system in the mobile application. As I said Xplore, 2021, pp. 194– 197
this mobile application is created as user-friendly and with [9]. M. H. M. R. Behshad Hosseinifard, "Classifying
user- friendly colors. Everyone can understand the word that depression patients and normal subjects using machine
we usedfor mobile applications. After all, we tried to make a learning techniques and nonlinear features from EEG
unique system for this problem. The solution came with this signal," Researchgate, October 2012, pp. 140–142
mobile application. Using this we are trying to avoid
depression amongSri Lankan software engineers.

V. CONCLUSION

Depression-enabler Most applications focus on treating

depression, anxiety, general mental health, stress, PTSD,
bipolar disorder, panic disorder, sleep disturbance,
schizophrenia, OCD, and addiction (non-drug and alcohol-
related addiction). Apps provide ways to boost mental health.
Relaxation, stress management, symptom monitoring,

IJISRT22NOV492 www.ijisrt.com 836