Skip to content
Conference Open access

An Explainable and Uncertainty-Aware Deep Learning Framework for Student Mental Health and Burnout Prediction

Sep 2026 · 2026 7th International Conference on Smart Electronics and Communication (ICOSEC) · pp. 1588-1594 · 0 citations · 18 references

TL;DR

This paper proposes an explainable and uncertainty-aware deep learning framework for student mental health and burnout prediction using structured behavioral and academic data and effectively integrates predictive accuracy, explainability, and uncertainty awareness within a lightweight architecture, making it suitable for educational mental health monitoring.

Abstract

The increasing prevalence of student mental health issues, including stress, anxiety, and burnout, highlights the need for reliable and scalable predictive systems for early intervention. Existing machine learning and deep learning approaches primarily focus on predictive accuracy while often lacking interpretability and uncertainty quantification, limiting their practical applicability. This paper proposes an explainable and uncertainty-aware deep learning framework for student mental health and burnout prediction using structured behavioral and academic data. An Artificial Neural Network (ANN) is employed for multiclass risk classification, while SHAP-based explainability provides global and local feature-level interpretations. Entropybased uncertainty estimation is incorporated to quantify prediction confidence and identify ambiguous cases, thereby improving decision reliability. Experimental results demonstrate an overall accuracy of 98.03%, with precision, recall, and F1-score values exceeding 0.97 across all classes. The model records only 118 misclassifications among 6000 test samples and achieves excellent ROC and precision-recall performance. SHAP analysis identifies burnout score, stress level, anxiety score, and depression score as the most influential predictors, while uncertainty analysis reveals an inverse relationship between prediction confidence and uncertainty. The proposed framework effectively integrates predictive accuracy, explainability, and uncertainty awareness within a lightweight architecture, making it suitable for educational mental health monitoring. Future work will focus on external validation using independent datasets and the integration of multimodal data to further improve the framework's generalizability and practical applicability. This version addresses the reviewers' concerns while keeping the abstract within a typical IEEE conference length.

Read PDF

Similar papers

Open access Sep 2026

A Software-Oriented Deep Learning Framework for Multiclass Academic Stress and Mental Health Risk Prediction

This study proposes a multiclass predictive framework for academic stress and mental health risk classification among students using a Kaggle dataset containing 25,000 records, and indicates that ensemble and deep learning approaches can effectively support multiclass mental health risk classification.

Shoaib Ahmad, A. Imran, Tahoona Kshif et al. · 0 citations
Open access Sep 2026

Explainable Machine Learning for Predicting Student Depression Risk

An explainable machine learning framework for predicting student depression risk using non-clinical demographic, academic, lifestyle, and psychosocial factors was developed and explainability analysis identified Suicidal Thoughts, Academic Pressure, and Financial Stress as the most influential predictors.

Reynalyn Cernechez, S. Hosseini, Muhammad Nadeem et al. · 0 citations
Open access Sep 2026

Early Depression Risk Screening using Leakage-Aware Ensemble Learning with Probability Calibration

A leakage-aware and calibration-aware ensemble learning framework for early depression risk screening among technical education students by excluding leakage-prone predictors during model development is presented, which provides an interpretable and methodologically transparent approach for depression risk prioritizati...

Mohitsinh Parmar, S. Pandya · 0 citations
Open access Nov 2026

Low-dimensional burnout risk classification in engineering students: the role of demand-recovery interaction features

Student burnout is a rapidly growing concern, with prior research linking its development to a wide range of psychological, behavioral, and contextual factors. Machine learning approaches have recently shown strong predictive performance for identifying burnout, but typically rely on high-dimensional datasets impract...

Arnav Saklani, Daisy Gohain, Kirankumar Manivannan · 0 citations
Open access 2026

Actionable and Trustworthy AI for Student Depression Prediction Using Multi-Level Explainability

It is critical to be able to identify and intervene with high-risk students as early as possible - this is most easily accomplished by developing scalable systems for providing such early warning signals. Although many researchers have developed highly accurate machine learning (ML) models to detect depression, poor cl...

S. Dash, Deepak Kumar Pradhan, M. A. Ahmed et al. · 0 citations
Open access Sep 2026

Construction of an Early Warning Model for Academic Burnout among High School Students Based on Learning Behaviour Data

An explainable machine-learning approach to identifying cases of academic burnout among high school students based on their observable learning behaviors is suggested, using a publicly accessible dataset that includes information about 649 adolescents and contains no validated burnout measurement.

Tian-Ge Xiang · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.