Publication

Abstract : This work evaluates the effectiveness of various machine learning (ML) and deep learning (DL) models in predicting and addressing student health concerns. The study compares the performance of multiple models, including traditional machine learning algorithms and modern deep learning architectures. A dataset with 30 samples and 13 psychological and emotional indicators was collected via a structured Google Form. Among the machine learning models, Random Forest demonstrated the highest accuracy of 85.52%, followed by Support Vector Machine (SVM) at 72.39%, Logistic Regression at 68.69%, and Gradient Boosting at 74.41%. On the deep learning front, the Feedforward Neural Network achieved an accuracy of 83.64 %, while Convolutional Neural Networks (CNN) outperformed other models with an accuracy of 87.12%. These findings underscore the potential of both machine learning and deep learning techniques in improving student health monitoring and prediction, offering valuable insights for future research and intervention strategies. The results further highlight that CNNs may be particularly effective in addressing complex health-related patterns among students.