Explainable AI-Driven Identification of Depression Levels for Early Mental Health Intervention in IT Students

Abstract

In today's rapidly evolving technological era, the IT industry has emerged as one of the most in-demand sectors among students. Students in IT faculties face significant academic pressure, leading to an increase in mental health problems. This research is related to the prediction of IT students' mental health status by means of machine learning (ML) approaches and explainable AI for early prevention and diagnosis. The proposed approach involves preprocessing the dataset, circumscribing various parameters of IT students and IT education. The primary objective of this research is to determine the factors that have a comparatively greater influence on the depression of students. Upon employing several ML models for outcome explanation, such as Random Forest Regressor, Gradient Boosting Regressor, AdaBoost Regressor, Linear Regression (LR), and Histogram-based Gradient Boosting Regressor, the best result was obtained from the LR with R² = 0.581867, mean squared error = 0.025011, and mean absolute error = 0.1197, which was later cross-verified using 10-fold results. After the cross-validation, LR obtained the mean R² value of 0.6008. Additionally, applying K-means clustering, a positive relationship was visualized between anxiety and depression, suggesting a proportional relationship between them. Lastly, Local Interpretable Model-agnostic Explanations and Shapley Additive Explanations were used for the outcome explanation to provide insights into the decisions of the models. These findings highlight the important aspects that influence the depression level of IT students, so that proper initiatives can be taken to improve the overall mental condition.

Received: 27 December 2025 | Revised: 18 March 2026 | Accepted: 8 May 2026

Conflicts of Interest

The authors declare that they have no conflicts of interest to this work.

Data Availability Statement

The data that support the findings of this study are publicly available in at:

1. https://www.kaggle.com/datasets/mohsenzergani/bangladeshi-university-students-mental-health
2. https://www.kaggle.com/datasets/abdullahashfaqvirk/student-mental-health-survey

Author Contribution Statement

Md. Rakeen Islam Nahin: Conceptualization, Methodology, Software, Formal analysis, Investigation, Data curation, Writing – original draft, Writing – review & editing, Visualization. Safiul Haque Chowdhury: Methodology, Validation, Writing – review & editing, Supervision. Mohammed Ibrahim Hussain: Supervision. Mohammad Minoar Hossain: Supervision. Mohammad Mamun: Supervision.