A Machine Learning-based system that predicts student academic performance and generates personalized learning recommendations to support data-driven educational decision-making.
Educational institutions often face challenges in identifying students who may struggle academically before their performance declines significantly. This project leverages Machine Learning techniques to analyze student-related attributes, predict performance levels, identify behavioral patterns, and provide personalized recommendations for academic improvement.
The system combines both supervised and unsupervised learning approaches to deliver predictive insights and tailored learning interventions through an interactive Streamlit application.
- Predicts student performance categories (High Performer, Average Performer, At-Risk Student)
- Identifies hidden patterns in student behavior using clustering techniques
- Generates personalized learning recommendations based on predicted outcomes
- Performs data preprocessing and feature engineering for improved model performance
- Provides real-time predictions through an interactive Streamlit interface
- Supports early identification of students requiring academic intervention
Used to classify students into performance categories based on historical academic and behavioral data.
Algorithm Used
- Random Forest Classifier
Used to discover natural groupings among students with similar learning and performance characteristics.
Algorithm Used
- K-Means Clustering
- Data Collection
- Data Cleaning
- Missing Value Handling
- Categorical Encoding
- Feature Engineering
- Model Training
- Model Evaluation
- Clustering Analysis
- Recommendation Generation
- Streamlit Deployment
The project derives meaningful features from raw student data, including:
- Attendance Percentage
- Academic Performance Indicators
- Study Habits
- Engagement Metrics
- Performance Ratios
- Behavioral Attributes
These engineered features improve predictive performance and enhance recommendation quality.
| Metric | Result |
|---|---|
| Classification Accuracy | 83–85% |
| Evaluation Metrics | Accuracy, Precision, Recall, F1-Score |
| Validation Approach | Train-Test Split and Cross Validation |
The model demonstrated reliable predictive performance and good generalization on unseen student data.
The recommendation system generates personalized suggestions based on predicted student performance and cluster assignment.
- Additional learning resources
- Focused revision plans
- Mentorship recommendations
- Attendance improvement guidance
- Structured study schedules
- Consistent revision strategies
- Performance improvement recommendations
- Advanced learning resources
- Challenge assignments
- Enrichment activities
- Python
- Pandas
- NumPy
- Scikit-Learn
- Matplotlib
- Seaborn
- Streamlit
- Random Forest Classification
- K-Means Clustering
- User enters student information.
- Data is preprocessed and transformed.
- Machine Learning model predicts student performance.
- Clustering module identifies student segment.
- Recommendation engine generates personalized suggestions.
- Results are displayed through the Streamlit dashboard.
Student-Performance-Recommendation-System/
│
├── app/
├── data/
├── screenshots/
├── src/
├── notebook/
└── README.md
- Deep Learning-based prediction models
- Explainable AI (XAI) integration
- Personalized recommendation ranking
- Student progress tracking dashboard
- Automated retraining pipeline
- Cloud deployment
Arjun Goel
MCA Candidate, Vellore Institute of Technology (2025–2027)
LinkedIn: linkedin.com/in/arjun11goel
GitHub: github.com/arjun11goel