🎓 Explainable LMS – Intelligent Student Performance Prediction System 📌 Overview
This project is a full-stack Learning Management System (LMS) integrated with an Explainable Machine Learning model to predict student academic performance.
The system allows:
🎥 Students to watch course videos
📚 Access learning resources
📝 Attend online tests
📊 Receive performance-based predictions
👩🏫 Teachers to view predictions and recommendations
The prediction model is built using XGBoost and achieves an accuracy of 89%.
🚀 Core Features 👨🎓 Student Features
User authentication (Login / Signup)
Course enrollment
Watch course videos
Access course resources
Attend online tests
View test results
Receive predicted performance (Pass / Fail)
👩🏫 Teacher Features
Create and manage courses
Upload videos & resources
Create and manage tests
View student submissions
Access student performance predictions
View ML-based recommendations
Analyze student risk levels
🧠 Machine Learning Module
The system includes a trained XGBoost Classifier that:
Predicts student academic outcome (Pass / Fail)
Uses engagement & performance features
Provides interpretable insights using SHAP
📊 Model Performance Metric Value Accuracy 89% Model XGBoost Classifier Type Binary Classification 🔍 Explainability with SHAP
To ensure transparency, the system integrates SHAP (SHapley Additive exPlanations):
Global feature importance visualization
SHAP summary plots
Individual student-level waterfall plots
Clear explanation of prediction reasoning
This enables teachers to understand why a student is predicted to pass or fail.
⚙️ Installation & Setup
Backend Setup python -m venv venv venv\Scripts\activate pip install -r requirements.txt
Run backend:
python app.py
Frontend Setup cd frontend npm install npm run dev
🎯 System Objectives
Improve academic performance prediction
Provide transparent AI-based insights
Support teachers in identifying at-risk students
Enhance digital learning environments
Integrate ML with real LMS workflows
🔮 Future Enhancements
Cloud deployment
Personalized adaptive learning paths
Analytics dashboard for administrators
👩💻 Contributors
Gayathri G Arya KJ Shehin T Shaji Asvin Thadevoos
📄 License
Developed for academic and research purposes.