Predictive Model for Assessing Student Mental Health Risk Based on Academic Data Using Machine Learning Algorithm

This study proposes a predictive model for assessing student mental health risk based on academic data using machine learning techniques. The objective is to identify the most effective classification algorithm for binary classification tasks within this context. Three classifiersDecision Tree, Random Forest, and Support Vector Classifier (SVC)-were evaluated using a structured machine learning pipeline. The pipeline included data preprocessing, feature selection, and hyperparameter tuning using GridSearchCV. Model performance was assessed using standard evaluation metrics: accuracy, precision, recall, and F1-score. Among the three models, the Random Forest classifier achieved the highest performance, with an accuracy of 73.98 %, and demonstrated a strong balance between precision and recall. These findings suggest that Random Forest is a suitable choice for similar classification problems in educational and psychological analytics. The study also discusses potential improvements using advanced ensemble methods and feature engineering to further enhance prediction effectiveness.