EduPredict: Academic Performance Analysis with Predictive Modelling

📌 Project Overview

This project explores predicting student academic performance using machine learning techniques. By analyzing behavioral and demographic data (e.g., gender, guardian relation, attendance), we identify key factors influencing student success. It includes a desktop app for interactive predictions.

Demo: https://igit.me/EduPredict

📊 Dataset

• Name: xAPI-Edu-Data
• Source: Kaggle (Open-access educational repository)
• Size: 480 students, 17 features
• Target Variable: Class (Low, Medium, High Performance)
• Key Features: Demographics (gender, nationality, relation to guardian—Mother vs. Father), behavioral (raised hands, resource visits, announcements, discussions, parental surveys, absences).

🛠 Data Preprocessing

• Data Cleaning: Handled missing values, standardized numerical features
• Feature Engineering: Encoded categorical data
• No Dimensionality Reduction: All 17 features retained for interpretability

🤖 Machine Learning Models Used

We applied seven classification algorithms:

Model	Purpose
📌 Decision Tree	Interpretability & feature importance
🌳 Random Forest	Improved accuracy & robustness
📈 Support Vector Machine (SVM)	Identified attendance threshold
📍 K-Nearest Neighbors (KNN)	Analyzed attendance & participation impact
📊 Naïve Bayes	Probabilistic classification
🚀 Gradient Boosting	High accuracy & feature importance
🧠 Artificial Neural Network (ANN)	Studied parental influence

🔍 Key Insights

✅ Attendance Matters: Students with >7 absences are likely low-performing.
✅ Participation Helps: More class participation (raised hands) → Higher performance.
✅ Parental Involvement is Crucial: Students whose parents answered school surveys performed better.
✅ Demographics: Explored gender and guardian relation (Mother vs. Father) impacts via statistical tests and visualizations.

🚀 How to Run the Project

1. Clone this repository:

   git clone https://github.com/ScriptSorcerer23/EduPredict.git
   cd EduPredict

2. Install dependencies (create requirements.txt if needed):

   pip install pandas numpy scikit-learn seaborn matplotlib tkinter imbalanced-learn

3. Run the analysis (Jupyter Notebook):

   jupyter notebook Academic_Performance_Project.ipynb

4. Run the Desktop App (e.g., for ANN predictions):

   python ann_front_end.py

   • Use individual front-end scripts (e.g., decision_tree_front_end.py) for model-specific GUIs.

📁 Project Structure

• Academic_Performance_Project.ipynb: Main analysis notebook with EDA, modeling, and visualizations.
• *_front_end.py: Desktop GUI scripts for each model.
• *.pkl: Trained models, preprocessors, and scalers.
• xAPI-Edu-Data.csv: Dataset.
• Data_Mining_Project_Presentation.pptx: Project slides.
• *.png: Model diagrams and plots.

📜 License

This project is open-source. Feel free to contribute or adapt! This project is open-source and available under the MIT License.