Related Experiment Video
Updated: Jul 10, 2025
06:37
Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention
Published on: December 15, 2023
3.8K
Impact of mobile connectivity on students' wellbeing: Detecting learners' depression using machine learning
Muntequa Imtiaz Siraji1, Ahnaf Akif Rahman1, Mirza Muntasir Nishat1
1Department of Electrical and Electronic Engineering, Islamic University of Technology, Gazipur, Dhaka, Bangladesh.
Plos One
|November 27, 2023
Summary
Student mental health is crucial, especially with increased mobile device use during the pandemic. Machine learning models can accurately identify depression levels in students, aiding early intervention and wellbeing support.
Area of Science:
- Psychology
- Computer Science
- Public Health
Background:
- The COVID-19 pandemic increased mobile device reliance, impacting student mental health.
- Blended learning models further accelerate this trend, necessitating mental health monitoring.
- Student vulnerability to depression is heightened during academic careers.
Purpose of the Study:
- To investigate the mental health of students concerning continuous mobile device use.
- To develop an automated system for identifying and classifying depression severity in students.
- To assess the effectiveness of machine learning algorithms and feature engineering techniques for depression detection.
Main Methods:
- A cross-sectional survey collected data from 444 university students.
- Eight machine learning algorithms were employed for depression identification and classification.
- Feature selection (Chi-square, RFE) and feature extraction (PCA, SparsePCA) methods were utilized.
Main Results:
- Machine learning models with feature engineering improved accuracy by 3-15%.
- SparsePCA with CatBoost classifier yielded optimal accuracy, F1-score, and ROC-AUC.
- 44% of students showed no depression, 25% mild-to-moderate, and 31% severe-to-extreme.
Conclusions:
- Machine learning models with feature engineering are effective for multi-stage depression detection in students.
- This automated system can aid in early identification and intervention for student mental health.
- The developed model has potential applications in other disciplines for depression screening.
