Leveraging AI for detection of mental health disorders: an investigation of multiple classifiers

Mental health disorders, particularly stress, anxiety, and depression, have become a significant global concern, affecting millions of individuals worldwide. Traditional diagnostic methods rely on self-reported assessments and clinical interviews, which are often subjective and time-consuming. Machine learning (ML) offers a promising alternative by automating mental health detection using data-driven approaches. This study investigates the effectiveness of four ML models—Neural Network, Support Vector Machine (SVM), K-Nearest Neighbors (KNN), and Random Forest—for predicting mental health conditions using the MHP Anxiety, Stress, Depression Dataset of University Students. The dataset was preprocessed through feature selection, categorical encoding, and normalization to optimize model performance. Experimental results demonstrate that SVM achieved the highest accuracy for stress detection and for anxiety classification as well, while the Neural Network model showed strong overall performance across all three conditions. Comparative analysis reveals that deep learning models outperform traditional classifiers in handling complex mental health assessment patterns. However, ethical concerns, data privacy issues, and the interpretability of ML models remain key challenges. Future work should focus on integrating multimodal data sources, explainable AI techniques, and real-time mental health monitoring solutions.