Analysing Student Depression through Cyber-Activity Patterns Using Heterogeneous Graph Attention Networks

This study focusses on the crucial matter of identifying student sadness by analysing patterns of cyberactivity. The technology utilizes Heterogeneous Graph Attention Networks (HGATs) to represent and examine intricate connections within the data. The dataset used for training and testing the model was obtained from Kaggle. The model achieved an accuracy of $\mathbf{9 6 . 8 \%}$, a precision of $\mathbf{9 5 . 4 \%}$, a recall of $94.7 \%$, and an AUC-ROC of $98.3 \%$. The findings of this study demonstrate a substantial improvement over current models, indicating the high efficacy of HGATs in accurately capturing complex patterns in cyber-activity data. The results indicate that this method can function as a potent instrument for identifying depression at an early stage, with wide-ranging implications for monitoring the mental health of students and implementing intervention techniques.