Multi-class adaptive learning for predicting student anxiety

Abstract

This project presents a machine learning-based approach for predicting student anxiety levels using a multi-class adaptive learning framework. The system analyzes various student-related attributes such as age, gender, course, academic performance, marital status, and mental health indicators like depression and panic attacks to identify different levels of anxiety, including no anxiety, mild, moderate, and severe anxiety. Data preprocessing techniques such as handling missing values, label encoding for categorical features, and normalization are applied to improve data quality and model performance. Multiple classification algorithms, including Decision Tree, Random Forest, and Support Vector Machine, are implemented and compared to achieve accurate predictions. The adaptive learning component enables the system to update and improve its performance as new data is introduced, making it more robust and scalable over time. A user-friendly web application is developed using Flask, allowing administrators to upload datasets, train models, and evaluate performance, while users can input their details to receive real-time anxiety predictions. This system aims to assist in early detection of mental health issues among students, enabling timely intervention and support, thereby contributing to improved academic performance and overall well-being.