Enhancing Student's Stress Level Prediction Using Active Learning and Dual Stack Approach

View at publisher Summarise with VIRAT
Authors: R. Shanthakumari, S. Sathya, S. Shri Ram, R. Sabarivasan, N. Yogeshwaran
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Stress among students significantly affects mental health, academic performance, and well-being, often leading to anxiety, depression, and burnout. Early detection of stress is essential for prevention and intervention. This research proposes an active learning-based dual stacking model, combining Gradient Boosting and SVM classifiers, to predict student stress levels. Feature selection uses Random Forest importance scores and RFE to focus on relevant features. Active learning selects uncertain samples to improve training efficiency and model generalization. The model, validated on a student dataset, achieved 99.09% accuracy, with strong precision, recall, and F1-scores, demonstrating its value in supporting student well-being through targeted stress prevention.

Fields of Study & Subjects

Online Learning and Analytics Artificial Intelligence in Healthcare Student Stress and Coping

Publication details

Journal / source
CRC Press eBooks
Pages
131–138
Publication date
26 March 2026
Language
EN
Document type
Book chapter