A Computational CNN-LSTM-Based Mental Health Consultation System in a College Environment

Abstract

In recent years, the mental health problems of college students have gradually attracted significant attention. Increasing academic pressure, employment competition, social adaptation challenges, and other stressors have led to a rise in mental health issues among college students, including anxiety, depression, and social disorders. These problems adversely affect students' academic performance, with long-term negative implications for their future career development. The COVID-19 pandemic further exacerbated psychological stress through isolation measures, distance learning, and social restrictions, highlighting the limitations of traditional mental health service models. In this study, we constructed a deep learning-based mental health counseling system tailored for college students. Utilizing a dataset of 467 valid questionnaires combined with students' behavioral data—such as library study time, internet usage habits, exercise frequency, and canteen consumption records—we employed Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks to analyze and predict students' mental health statuses. Additionally, natural language processing (NLP) techniques were integrated to provide personalized psychological support and interventions. The combined CNN-LSTM model achieved an overall accuracy of 85.7%, a mean squared error (MSE) of 0.041, and an F1 score of 85.2%, demonstrating high precision in identifying and predicting various mental states, including anxiety and depression. Furthermore, trend analysis over six months revealed significant insights into the dynamics of students' mental health, such as increasing stress and declining emotional stability towards the end of the semester. User feedback indicated an 82.6% satisfaction rate with the system's professional content and 80.1% with emotional support, while 77.5% expressed overall satisfaction. These results confirm that the proposed system effectively enhances mental health monitoring and intervention, providing robust technical support for college mental health services and offering a scalable solution for large-scale applications.