A PREDICTIVE AND PREVENTIVE MACHINE LEARNING FRAMEWORK FOR STUDENT DROPOUT ANALYSIS AND COUNSELING

Student dropout is a critical issue in educational institutions that affects academic performance,institutional reputation, and student career development. Early identification of at-risk students and timely intervention are essential to reduce dropout rates and improve student retention. This paper presents a web-based Student Dropout Prediction and Counseling System developed using machine learning and intelligent counseling support. The proposed system operates in two modes: student mode and faculty mode. In student mode, learners can perform self-evaluation by entering academic attributes such as attendance, marks, and behavior to assess their dropout risk. In faculty mode, teachers can upload class datasets and predict students who are at risk of dropping out.The system utilizes a Random Forest classifier trained on key academic and behavioral features to predict dropout risk effectively. The trained model is integrated into a Flask-based web application that dynamically processes student data and generates risk predictions. In addition to prediction, the system provides dynamic counseling support through AI-based chat, notification alerts, and video counseling features,enabling both preventive and corrective intervention. The integration of prediction,analysis, and real-time counseling distinguishes the proposed system from traditional prediction only approaches. Experimental evaluation demonstrates that the system can accurately identify at-risk students and support timely academic guidance, making it suitable for real world educational environments.