A Novel ensemble Machine Learning Algorithm for detection of Health Early Warning for College Students

The rising prevalence of mental and physical health issues among college students has become a critical concern, emphasizing the need for intelligent health monitoring systems capable of identifying potential risks at an early stage. Academic pressure, social stress, lack of sleep, and unhealthy lifestyle choices contribute significantly to deteriorating student health, which can impact academic performance and overall well-being. To address this, the present study recommends a novel ensemble machine learning framework that integrates three powerful classifiers: Support Vector Machine (SVM), Random Forest (RF), and Extreme Learning Machine (ELM). This hybrid approach aims to elevate the accuracy and reliability of health early warning detection systems. The proposed model is trained and examined by utilizing the publicly available Student-Health-Data dataset, which contains a wide range of physiological, psychological, and behavioural indicators relevant to student wellness. The dataset underwent thorough preprocessing, including normalization, handling of missing values, and feature selection, to improve learning quality and reduce noise. The ensemble method exploits the decision-making precision of SVM, the robust feature-handling capacity of RF, and the fast-learning ability of ELM. Experimental results revealed impressive performance metrics: 97.8% accuracy, 97.5% precision, 97.1% recall, 97.3% F1-score, and an AUC-ROC score of 0.986. These outcomes demonstrate the proposed model's potential in accurately identifying health risks and enabling educational institutions to initiate timely and proactive interventions for student wellbeing.