ACADEMICPREDICTOR

Multiclass classification system to predict academic success of university students using XGBoost and Random Forest

The Problem

Universities face high student dropout rates that affect both educational institutions and students. Without early predictive tools, it is difficult to identify at-risk students and provide necessary support before they drop out. This situation generates economic losses for institutions and personal frustrations for students.

The Solution

Development of a predictive system based on machine learning that analyzes multiple academic and socioeconomic factors to identify risk patterns. The system uses XGBoost and Random Forest algorithms with 85% accuracy, enabling early and personalized interventions. The web interface facilitates access to real-time predictions.

Impact

Enables educational institutions to implement proactive retention strategies, reduce dropout rates, and improve student academic success through early and targeted interventions.