Empowering education through data! EduRisk is a machine learning project that predicts underperforming students and provides actionable insights to improve academic outcomes. It combines data preprocessing, feature engineering, and predictive modeling to identify at-risk learners early.
Leverage advanced ML algorithms to predict students who are likely to underperform based on academic and behavioral data.
Comprehensive charts and graphs to analyze performance trends and model metrics.
Includes AdaBoost, CatBoost, and XGBoost with evaluation metrics like Accuracy, Precision, and F1-score.
Compare models and datasets visually to choose the best-performing approach for your scenario.
Export results, graphs, and insights for reporting and presentations.
Here’s a snapshot of the best test-set metrics across different datasets:
-
Mathematics Dataset → Best Model: AdaBoost
✅ Accuracy: 0.696 | ✅ Precision: 0.692 | ✅ F1: 0.684 -
Portuguese Dataset → Best Model: CatBoost
✅ Accuracy: 0.708 | ✅ Precision: 0.674 | ✅ F1: 0.662 -
Exams Dataset → Best Model: XGBoost
✅ Accuracy: 0.485 | ✅ Precision: 0.484 | ✅ F1: 0.483
Clone this repository to your local machine to get started with EduRisk:
git clone https://github.com/your-username/EduRisk.git