Authors: Bharadwaj Patil, Mohammed Zaid Pathan, Om Devkar, Professor Nilesh Ahire
Abstract: Student dropout is a critical challenge in educational institutions worldwide, resulting in signifi-cant social, economic, and academic consequences. This paper presents an AI-Based Drop-Out Prediction and Counselling System that leverages machine learning algorithms to proactively identify students who are at danger and offer timely automated counselling inteventions. The suggested system combines several data sources.— including academic performance, attend-ance records, socio-economic indicators, and behavioral patterns — to build predictive models using algorithms such as Random Forest, Gradient Boosting (XGBoost), Support Vector Ma-chine (SVM), Logistic Regression, and an Artificial Neural Network (ANN). The system achieves a prediction accuracy of 94.7%, a precision of 93.2%, recall of 95.1%, and an F1-score of 94.1% on the validation dataset. An intelligent counselling module is also designed to provide personalized, AI-driven recommendations to students flagged as high risk. Experimental results on a dataset of 5,000 student records demonstrate the superiority of the suggested strategy over existing baseline methods. The system is designed as a web-based platform accessible to adminis-trators, faculty, and counsellors, enabling real-time monitoring and intervention
