Insightful Mental Health Tool for Students

Postolache Amalia Maria, Ioan Daniel Pop

and Adriana Mihaela Coroiu

Department of Computer Science, “Babes-Bolyai” University, 400084, Cluj-Napoca, Romania

Keywords:

Students’ Mental Health, Multilayer Perceptron, Artiﬁcial Neural Network, Intelligent Data Analysis.

Abstract:

This paper aims to understand the nexus between academic context and student well-being. It provides an

overview of the prevalence of mental health problems among students and the ways in which these issues can

impact academic performance. Nevertheless, the factors of the academic environment that can inﬂuence a

student’s mental health are also discussed, outlining the importance of addressing such issues in the student

community and the potential long-term impact it can have on individuals, leading us to the reasoning behind

creating the application that serves as the focus of this paper. Considering the practical use of the application

and what it offers, this paper will also discuss the intelligent analysis of the collected data, enabling further

interpretation by mental health professionals.

1 INTRODUCTION

Living in a world where the professional success of a

high school graduate can be predicted by his choice

to pursue higher education studies and the career

prospect is directly impacted by his academic perfor-

mance, we were determined to ﬁnd a way of support-

ing the students who do not always meet the expec-

tations of either their families or their educators. A

student who drops out of school embodies the image

of incompetence or disinterest most of the time, but

how often have people taken the time to analyze the

factors that led to the dropout or the poor academic

performance that often precedes it?

A student’s academic functioning and, subse-

quently, career outlook are unfortunately heavily in-

ﬂuenced by mental health disorders, given that most

factors weighing in on a student’s mental health state

originate in the academic environment. These factors

embody academic expectations and pressure to suc-

ceed, meet deadlines, manage time efﬁciently, or even

cover the cost of education.

Mental health is an important aspect of overall

well-being, and it can be particularly challenging for

students to maintain good mental health due to the

many stresses and demands of the academic environ-

ment. In recent years, college student populations

have witnessed a surge in symptoms of depression,

anxiety, and other mental illnesses, accompanied by

https://orcid.org/0000-0002-3740-6579

https://orcid.org/0000-0001-5275-3432

a consistent growth in the demand for counseling ser-

vices (Duffy et al., 2019). College students frequently

experience mental health issues, which are clearly

connected with lower academic performance. Addi-

tional research is required to determine whether this

link could be causal and, if so, whether therapies

aimed at addressing mental health problems could en-

hance academic performance.

The college years are a critical developmental

stage in which students transition from late adoles-

cence to emerging adulthood (Arnett, 2023). Some of

the long-term adverse outcomes (including persistent

emotional and physical health problems or relation-

ship dysfunction) may be inﬂuenced by mental health

issues that arise throughout the college years, as these

years are a peak phase for the emergence of a wide

spectrum of mental disorders.

According to research conducted by the Roma-

nian National Council for Higher Education Financ-

ing in 2018, around 74% of Romanian high school

graduates pursued their studies in higher education.

A study by the Romanian Ministry of Education re-

vealed that, in the 2019-2020 academic year, the over-

all dropout rate for Romanian universities was 10.4%,

which was a slight decrease from the previous year’s

rate of 10.5%. Previous research has found that col-

lege students with mental disorders or illnesses are

twice as likely to drop out without obtaining a de-

gree. There are fewer studies that focus on the associ-

ation between mental health problems and academic

success in college. The majority of data supports the

808

Postolache, A. M., Pop, I. D. and Coroiu, A. M.

Insightful Mental Health Tool for Students.

DOI: 10.5220/0013357800003932

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 17th International Conference on Computer Supported Education (CSEDU 2025) - Volume 2, pages 808-814

ISBN: 978-989-758-746-7; ISSN: 2184-5026

conclusion that depression and suicidal thoughts and

behaviors are associated with a lower grade point av-

erage (De Luca et al., 2016).

In this study we analyze other relevant works from

the specialized literature and present the creation of

an application for identifying mental problems among

students using an intelligent data analysis.

2 THE LITERATURE REVIEW

2.1 Educational Context and Mental

Health

The article (Bruffaerts et al., 2018) published in the

Journal of Affective Disorders investigates the link

between mental health problems and academic perfor-

mance among ﬁrst-year college students. It is carried

out as part of the World Mental Health Surveys In-

ternational College Student (WMH-ICS) project and

focuses on students at KU Leuven University in Bel-

gium. The primary objective is to identify the impact

these issues have on academic performance, as mea-

sured by Academic Year Percentage (AYP) and grade

point average (GPA).

The study involved a sample of 4,921 ﬁrst-year

students who completed an online mental health sur-

vey. Administrative data were used to assess aca-

demic achievement at the end of the year as well as

demographic information. Mental health problems

were assessed using the Global Appraisal of Indi-

vidual Needs Short Screener (GAIN-SS), a validated

instrument for screening emotional and behavioral

problems.

Among the main results was that about 34.9% of

students reported mental health problems in the past

year. The most common problems reported are those

of an internalized nature such as depression or anx-

iety, being present in 23.70% of those who reported

problems. Hyperactivity, impulsivity or other types

of externalized problems were present in 18.30% of

the students. Problems such as substance use or anti-

social behavior were less frequent, 5.40% and 0.10%

respectively.

The data show that mental health problems have

a signiﬁcant negative impact on academic perfor-

mance, with students with internalizing and external-

izing problems experiencing a 2.9% to 4.7% decrease

in AYP, corresponding to a 0.2–0.3 decrease in GPA

(Bruffaerts et al., 2018). The effect of externaliz-

ing problems was more pronounced in departments

with lower overall academic performance. This sug-

gests that lower performing academic environments

may amplify the impact of mental health problems on

school outcomes.

Of course, there are some limitations regarding

the (Bruffaerts et al., 2018) study, such as the small

number of cases of students with antisocial behavior

within the sample, or the lack of data on academic

performance in high school, which could have inﬂu-

enced the observed relationships. Another thing that

could have limited the study is the fact that the data

are based on a single screening tool and refer to a sin-

gle university, limiting the generalization of the re-

sults. This study indicates the need for interventions

at the university level. Supportive mental health inter-

ventions could help improve academic performance

and reduce the risk of dropping out.

Another work that addresses the connection be-

tween mental health and the academic performance

of students in the ﬁrst year of college is (Wyatt et al.,

2017). The authors used data from the American Col-

lege Health Association-National College Health As-

sessment (ACHA-NCHA) and underline the fact that

there is a strong correlation between the mental health

of students and the drop in academic performance and

the risk of dropping out.

The data set on which the experiments were per-

formed were collected from students in 2011 and con-

sists of 66,159 records. Forms were analyzed regard-

ing problems such as anxiety, depression, self-harm

and suicidal ideas and how they affect academic per-

formance in the ﬁrst year of college. According to the

authors the ﬁrst year of college in the United States

brings many changes in the lives of young people, be-

ing the period with the highest incidence of mental

health problems, with students frequently reporting

symptoms of depression and anxiety. Diagnoses in-

crease progressively as students advance through their

years of study, and female students are more prone to

such problems than their male peers. Depression and

anxiety are frequently associated with lower grades,

dropping out of classes and, in some cases, dropping

out of school.

The study makes certain recommendations for the

university, such as the inclusion of emotional edu-

cation classes and stress management techniques or

improving access through various methods (eg online

communication). The conclusions of the (Wyatt et al.,

2017) study show that promoting awareness and early

interventions in mental health is essential both in the

ﬁrst year of college and throughout the years.

A very useful study focused on the problem of

mental health and the relationship between it and aca-

demic performance is the paper (Zhang et al., 2024).

The authors of the study emphasize stress, depression

and anxiety in the post-pandemic period and their in-

Insightful Mental Health Tool for Students

809

ﬂuence on students. The sample used in the study

contains 600 students divided into two categories (300

ﬁrst year students and 300 ﬁfth year students). The

DASS (Depression, Anxiety, and Stress Scale) and

PHQ-9 (Patient Health Questionnaire-9) scales were

used to measure mental health, and academic perfor-

mance was assessed on a 12-point scale.

The obtained results showed that the ﬁrst-year stu-

dents have a higher level of stress (27.1 compared

to 24.2 for the 5th year), but they obtained higher

average academic scores (11.2 compared to 10.5).

The results indicate a negative correlation between

stress and academic performance (r = -0.25). An-

other impressive result is that anxiety and depression

were observed to have a signiﬁcant inﬂuence on aca-

demic performance (depression: r = -0.20; anxiety:

r = -0.18). The reduction in symptoms led to an

improvement in academic performance. The (Zhang

et al., 2024) study highlights the fact that students re-

ported a signiﬁcant impact of the pandemic on mental

health, with persistent symptoms of social isolation

and changes in the learning process. About 75% of

students reported difﬁculty adjusting to online learn-

ing, and 30% reported long-lasting physical symp-

toms after being infected with COVID-19. The study

highlights the need for preventive measures for stu-

dents’ mental health and shows that supportive inter-

ventions can help improve academic performance and

overall well-being.

2.2 Machine Learning in Students’

Mental Health

Considering the importance of mental health in the

lives of students and beyond, it goes without saying

that there are numerous studies that investigate these

correlations using machine learning and artiﬁcial in-

telligence techniques. Paper (Shaﬁee and Mutalib,

2020) presents the rise of mental health problems

among Malaysian higher education students as part

of a global phenomenon. These problems seriously

affect the daily life of students and their social inter-

actions, negatively inﬂuencing their academic perfor-

mance and their integration into the university envi-

ronment. In Malaysia, the Ministry of Higher Educa-

tion supports approximately 20 public institutions and

447 private institutions, which educate over 550,000

students annually. In this context, students face ma-

jor pressures such as separation from family, ﬁnancial

insecurity and uncertainties about their future careers,

all of which contribute to a high risk for mental health

problems. The study found that many students avoid

seeking help because of stigma or the belief that these

problems are normal in university life.

Among the types of mental health problems that

have been investigated are anxiety disorders, depres-

sive disorders and contributing factors. In the study

(Shaﬁee and Mutalib, 2020), several algorithms were

used for the analysis of mental health data, such as

Support Vector Machine and Neural Networks for the

classiﬁcation and analysis of psychometric responses,

and for the predictions related to the emotional and

mental states of the students, Random Forest and Lo-

gistic Regression were used.

The authors suggest that machine learning tech-

niques can play a crucial role in identifying and moni-

toring mental health among students, helping to detect

risks early and providing universities with guidance

for developing personalized support programs. In this

study, the authors obtained notable results regarding

the ability of machine learning models to predict stu-

dents’ mental health problems. SVM (Support Vector

Machine) and Neural Network models have demon-

strated high accuracy with accuracy rates over 70%-

96%, outperforming other methods in analyzing anx-

iety and depression data. These models have proven

effective in differentiating complex mental states and

have shown that factors such as lack of social sup-

port and academic stress are the most predictive of

students’ mental health.

The authors of the (Baba and Bunji, 2023) article

explore the application of a machine learning model

to identify mental health problems among college stu-

dents using data from annually collected health ques-

tionnaires. The study was conducted on a sample

of 3.561 students at a university in Japan, with re-

sponses including demographic variables, questions

about university life, and response time data. The

main aim of the study was to develop a prediction

model for identifying the risk of mental health prob-

lems, both within one year and in the following year,

based on students’ responses to health questionnaires.

The model was based on machine learning algo-

rithms, including LightGBM (Gradient Boosting Ma-

chine), which was compared with other models, such

as logistic regression and Random Forest. The cho-

sen model, LightGBM, demonstrated superior perfor-

mance as assessed by performance indicators such as

the Matthews Correlation Coefﬁcient (MCC), show-

ing a high ability to predict mental health problems.

Comparing various models, LightGBM was the

best performer in the study (Baba and Bunji, 2023),

especially in analyzing current year predictions and

anticipating problems for the next year. Answers to

questions about difﬁculties in university life (eg, anx-

iety about the future) were among the strongest pre-

dictors, with a high impact (gain of 0.131 - 0.216 and

SHAP values of 0.018 - 0.028). Variables related to

CSEDU 2025 - 17th International Conference on Computer Supported Education

810

response time, although considered relevant, did not

signiﬁcantly improve the performance of the model.

The authors of (Madububambachu et al., 2024)

study carry out a systematic analysis of machine

learning techniques applied in the diagnosis of men-

tal health disorders, with an emphasis on univer-

sity students. The study includes 30 papers pub-

lished between 2011 and 2024 that explore Convo-

lutional Neural Networks, Support Vector Machine,

Random Forest and deep neural networks algorithms

for diagnosing conditions such as depression, anxiety,

PTSD, ADHD and schizophrenia. The study applies a

PRISMA methodology to collect and analyze relevant

articles. Inclusion criteria were based on keywords

such as ”deep learning” and ”predict mental health,”

searching for relevant papers in recognized databases,

including IEEE Xplore and ScienceDirect. The ana-

lyzed data comes from various sources, such as fMRI

images, EEG signals, text data from social media and

medical surveys.

The study stands out for obtaining satisfactory

performances. The CNN and Random Forest mod-

els performed highly, with up to 99% accuracy, in di-

agnosing depression and anxiety. Social media text

and EEG analysis have proven to be effective tools

for prediction. The application of CNN and RF algo-

rithms on GPS and EEG data achieved an accuracy of

80%-99%, depending on the speciﬁcs of the dataset,

and the SVM algorithm demonstrated high efﬁciency

in detecting brain structural changes through imaging

data, reaching AUC (Area under the ROC Curve) of

up to 0.93.

The authors conclude the paper (Madububam-

bachu et al., 2024) that although machine learning

models show promise for diagnosing mental disor-

ders, there are important limitations, such as lim-

ited access to large datasets and the interpretability of

complex models such as neural networks. The study

suggests that standardization and the use of more di-

verse and longitudinal data sets could improve the

accuracy of diagnoses and the ease of implementing

these models in clinical practice.

The article (Rahman and Kohli, 2024) explores

the mental health issues facing international students,

using machine learning techniques to analyze the de-

mographic, cultural and psychosocial factors that in-

ﬂuence this category. The main aim was to create

a predictive model based on machine learning that

could identify the risk of depression among interna-

tional students in the UK. This approach is motivated

by increasing cases of depression and anxiety among

international students, often associated with academic

stress, ﬁnancial difﬁculties, culture shock and other

adjustment problems.

The research was based on two unique data sets,

the ﬁrst obtained through a survey applied to a group

of 87 international students between February and

March 2023 used for the training part of the model,

and the second used for testing contains data from

201 international students and was used to train and

test depression prediction models. The second set

is known as A Dataset of Students’ Mental Health

and Help-Seeking Behaviors in a Multicultural En-

vironment. In the study (Rahman and Kohli, 2024),

four machine learning algorithms were used: Logis-

tic Regression, Decision Tree, Random Forest and K-

Nearest Neighbors. The performance evaluation of

each model was done by accuracy, sensitivity, speci-

ﬁcity, precision, and AU-ROC curve metrics.

The results obtained emphasize the importance

of speciﬁc demographic and psychosocial factors in

the prediction of depression. The main contribu-

tors to international students’ mental health problems

were found to include ﬁnancial difﬁculties, academic

stress, homesickness, isolation, and culture shock.

Speciﬁcally, students in the 21-25 and 26-30 age cat-

egories, women and singles are more prone to depres-

sion and anxiety. The results also show that students

who have little social support or feel excluded from

the community have a higher risk of mental health

problems. In terms of algorithm efﬁciency, the Ran-

dom Forest model had the highest accuracy of 80%,

demonstrating a better ability to identify depressed

students. It had a high speciﬁcity in correctly iden-

tifying cases without depression, but a lower sensitiv-

ity in identifying positive cases, suggesting a stronger

performance in conﬁrming mental well-being than in

detecting the risk of depression. Although the results

obtained in (Rahman and Kohli, 2024) are very good,

a major limitation of the study is the relatively small

size of the samples, both of the primary and secondary

sets.

A paper aiming to detect and predict mental health

disorders among students by applying various ma-

chine learning techniques is (Sahu and Debbarma,

2022). The authors tested the performance of the al-

gorithms: Logistic regression, decision trees, random

forest, closest neighbors k, and neural network to pre-

dict the risk of mental disorders among young peo-

ple. In the study a data set with 5,840 records was

used. To evaluate the models the authors used the

confusion matrix and calculated the accuracy, preci-

sion, sensitivity and AUC-ROC. The Neural Network

model demonstrated the highest accuracy of 99.03%,

being the best performer in predicting the risk of men-

tal disorders in students. The random forest and K-

nearest neighbor algorithms also scored highly, sug-

gesting their effectiveness in mental health assess-

Insightful Mental Health Tool for Students

811

ment. The Neural Network model had an AUC score

of 0.98, indicating excellent performance in identify-

ing differences between positive and negative cases

of mental disorder risk. This article (Sahu and Deb-

barma, 2022) contributes to the literature by providing

an automated and integrated solution to predict stu-

dent mental health using machine learning algorithms

in a way that is accessible to educational institutions.

3 MENTAL HEALTH CHECKER

FOR STUDENTS

As part of this research, we have created a practical

application that aims to check the mental health of fe-

male students. Mental Health Checker for Students

is a web application whose usability and functionality

revolve around a mental health quiz that analyzes how

the academic environment can impact students’ daily

activities. Through our research and application of

our ﬁeld-related knowledge, we have created a power-

ful tool for educators and mental health professionals

to better understand and support students in achieving

their full potential.

3.1 Concept

The concept of the application is that the student can

use the safety of anonymity to expose his vulnerabil-

ities related to the academic context. Given this, the

features responsible for user management were built

so that the completion of the quiz is in no way de-

pendent on or related to registration within the appli-

cation. When entering the web site, the user is wel-

comed with the landing page of the application, and

it gives the user the possibility to be redirected to all

the available resources within the app, either by using

the navigation bar or from the slideshow containing a

presentation of the functionalities.

Through the completion of the quiz by users, the

aim of investigating the extent to which mental health

problems are associated with academic functioning

can be met. Additionally, the application provides de-

tailed analytics on user activity, including quiz com-

pletion rates, quiz scores, and user engagement, al-

lowing users to track their progress in improving their

mental health state and subscribe to a newsletter that

provides regular self-care tips via mail. With person-

alized user proﬁles, individuals can view their most

recent quiz score, update subscription preferences or

personal information, and monitor their progress over

time. Moreover, the administrator user role is avail-

able for managing user accounts and newsletter dis-

tribution, with the admin being able to send the user

the weekly or monthly newsletter (depending on the

user’s preference) or delete the user’s account.

3.2 Process of Creating the Quiz

When preparing the dataset for creating the mental

health quiz for the web application, several impor-

tant considerations were taken into account, such as

legitimate research, the topics’ diversity, reliability,

and adaptability. The questions included in the men-

tal health quiz were carefully curated based on estab-

lished research and with the help of a professional,

a certiﬁed psychologist with over 15 years of work

experience. This ensures the validity and reliability

of the quiz in measuring relevant aspects of mental

health within the academic context. The questions

cover a wide variety of topics related to mental health,

including stress, anxiety, depression, time manage-

ment, coping mechanisms, and social support. By

including a diverse set of topics, the quiz aims to

provide a thorough assessment of a student’s men-

tal well-being within the academic environment. The

dataset was created with adaptability and scalability

in mind. It can be easily modiﬁed or expanded to

accommodate future updates, improvements, or cus-

tomization based on feedback or emerging research

in the ﬁeld of mental health. Regarding the sensitiv-

ity of the questions and the privacy of the users taking

the quiz, steps were taken to ensure that the questions

were respectful and non-intrusive, and along with the

option of taking the quiz without registration, the con-

ﬁdentiality and anonymity of the participants were

protected.

3.3 Questions Dataset

The quiz consists of the following questions:

• About how often did academic activities (courses,

laboratories, exams) make you feel stressed

(tense, irritable)?

• About how often did you feel that your effective-

ness in learning is affected by the pressure that is

placed on you?

• About how often was your sleep schedule affected

by stress during a year of study that involves at-

tending classes and laboratories?

• About how often have you struggled with anxi-

ety when you had to present your work in front of

your peers or teachers?

• About how often did you fail to meet the deadlines

for handing in assignments?

CSEDU 2025 - 17th International Conference on Computer Supported Education

812

• About how often do you think you have not been

able to maintain a healthy relationship with your

family because of the academic environment?

• About how often did you regret your choice to at-

tend college?

• About how often did you have a positive attitude

when encouraging others to attend college?

• About how often did you feel that everything was

an effort?

• About how often did you feel panicked or over-

whelmed by things in your life?

Regarding responses, the quiz’s answer scheme

includes the following choices:

• None of the time

• A little of the time

• Some of the time

• Most of the time

• All of the time

The ﬁve gradations offer a range of options that cover

a spectrum from no occurrence to a complete occur-

rence, creating a 4-scale questionnaire and providing

a detailed framework for assessing the impact of the

academic context on various aspects of mental health.

Given that this scheme is ﬂexible and adaptable to dif-

ferent scenarios or contexts within the academic en-

vironment, it can also be easily modiﬁed or expanded

upon to accommodate speciﬁc needs or changes in the

ﬁeld.

3.4 Intelligent Data Analysis

For an intelligent analysis of the data gathered with

the help of the application, and more speciﬁcally, in

order to obtain a prediction for the next score on the

mental health quiz, we used the multilayer perceptron

classiﬁer model to manipulate the data extracted from

the StatisticalEntry table from the database, which

contains information about every quiz completion un-

til the present. The training data (the data extracted

from the StatisticalEntry table) is propagated to the

MLP through the input layers, then passes through the

hidden layers. Next, by applying the activation func-

tion (in this speciﬁc case, the Rectiﬁed Linear Unit

function), the output is generated at the output layer.

The predicted output will then be compared to the ac-

tual output; hence error will be calculated. Based on

the test dataset, we were able to make a prediction re-

garding the evaluation of the next possible score on

the quiz. The model accuracy, precision, recall, and

F1 score, also known as the harmonic mean of the

precision and recall, were also assessed. This exam-

ple of intelligent data analysis that performs predic-

tions can be a way to facilitate the interpretation or

further development of the data generated within the

application by mental health specialists.

Table 1: Statistical Metrics.

Performance Metric value

Accuracy 0.83

Precision 0.83

Reacll 1.00

F1 Score 0.91

The results obtained are very satisfactory and

shows that machine learning algorithms can be suc-

cessfully used in the early prediction of possible men-

tal problems of students. The results can be viewed in

Table 1.

4 CONCLUSIONS

Mental health is one of the most precious things we

have. Daily stress and problems in everyday life can

lead to serious mental health problems, conditions

such as depression, anxiety and suicidal thoughts are

more and more common among people, especially

among young people. The vast majority of young

people who face such problems are students, and their

early detection could lead to efﬁcient and quick treat-

ment.

Given that the application’s main focus is the men-

tal health quiz designed for students, further devel-

opment or use of the Mental Health Checker can in-

clude creating several quizzes targeting different as-

pects of mental health. Providing users with multi-

ple self-assessment tests can be an additional step to-

wards identifying struggles and getting the right help.

In terms of adding more functionalities to the appli-

cation, an online forum can also be created as a safe

place for students to express their thoughts and con-

cerns, with the possibility of receiving direct support

from a professional. This would also include creating

a new user role or assuming that behind the Adminis-

trator role is a certiﬁed psychologist. Moreover, as we

discussed in the previous chapter, the intelligent anal-

ysis of the collected data can be taken to an extent

where specialists in the mental health ﬁeld interpret

it, taking into account the relevant improvements that

can be made as a result of the analysis. At the same

time, we could apply other intelligent algorithms re-

garding the prediction of the next score on the mental

health quiz.

Insightful Mental Health Tool for Students

813

Mental Health Checker is addressed speciﬁcally

to students and it It provides a self-assessment process

for identifying mental health struggles, more impor-

tant It offers quick links and contact options to coun-

selors, professionals, or emergency lines.

We are aware of the limitations of this study, but

we consider it a good starting point for other research

in the ﬁeld of students’ mental health.

REFERENCES

Arnett, J. J. (2023). Emerging adulthood: The winding road

from the late teens through the twenties. Oxford Uni-

versity Press.

Baba, A. and Bunji, K. (2023). Prediction of mental health

problem using annual student health survey: machine

learning approach. JMIR Mental Health, 10:e42420.

Bruffaerts, R., Mortier, P., Kiekens, G., Auerbach, R. P.,

Cuijpers, P., Demyttenaere, K., Green, J. G., Nock,

M. K., and Kessler, R. C. (2018). Mental health prob-

lems in college freshmen: Prevalence and academic

functioning. Journal of affective disorders, 225:97–

103.

De Luca, S. M., Franklin, C., Yueqi, Y., Johnson, S.,

and Brownson, C. (2016). The relationship between

suicide ideation, behavioral health, and college aca-

demic performance. Community mental health jour-

nal, 52:534–540.

Duffy, M. E., Twenge, J. M., and Joiner, T. E. (2019).

Trends in mood and anxiety symptoms and suicide-

related outcomes among us undergraduates, 2007–

2018: Evidence from two national surveys. Journal

of Adolescent Health, 65(5):590–598.

Madububambachu, U., Ukpebor, A., and Ihezue, U. (2024).

Machine learning techniques to predict mental health

diagnoses: A systematic literature review. Clinical

Practice and Epidemiology in Mental Health: CP &

EMH, 20:e17450179315688.

Rahman, M. A. and Kohli, T. (2024). Mental health anal-

ysis of international students using machine learning

techniques. Plos one, 19(6):e0304132.

Sahu, S. and Debbarma, T. (2022). Mental health prediction

among students using machine learning techniques. In

International Conference on Frontiers of Intelligent

Computing: Theory and Applications, pages 529–

541. Springer.

Shaﬁee, N. S. M. and Mutalib, S. (2020). Prediction of

mental health problems among higher education stu-

dent using machine learning. International Journal of

Education and Management Engineering, 10(6):1.

Wyatt, T. J., Oswalt, S. B., and Ochoa, Y. (2017). Mental

health and academic performance of ﬁrst-year college

students. International Journal of Higher Education,

6(3):178–187.

Zhang, J., Peng, C., and Chen, C. (2024). Mental health and

academic performance of college students: Knowl-

edge in the ﬁeld of mental health, self-control, and

learning in college. Acta Psychologica, 248:104351.

CSEDU 2025 - 17th International Conference on Computer Supported Education

814