Student Performance Prediction with Explainable AI & Fairness

Welcome to the central repository for our final year thesis research. This project focuses on predicting student academic performance using Machine Learning and Deep Learning (LSTM, Diffusion Models), while emphasizing model explainability (SHAP, LIME) and fairness.

📂 Repository Navigation

Select a directory below to navigate to the specific workspace:

Directory	Description	Status
💻 CodeBase	The Main Project. Contains the cleaned, production-ready code, datasets, and final models used for the thesis results.	🟢 Active
📚 Literature & Sandbox	Research Hub. Contains literature reviews, PDF papers, and individual contributor workspaces (Istiaq, Maria, Nazme) for experiments.	🟡 Ongoing
Thesis Paper	The Manuscript. Contains the LaTeX files, images, and bibliography for the final research paper.	🔴 Draft

---

🚀 Getting Started (Main Codebase)

To run the main project code, please follow these steps:

1. Clone the repository:

   git clone [https://github.com/your-username/your-repo-name.git](https://github.com/your-username/your-repo-name.git)

2. Navigate to the CodeBase:

   cd CodeBase

3. Install Dependencies:

   pip install -r requirements.txt

4. Run the Analysis:
   • Open src/student_performance_prediction.ipynb to view the primary prediction pipeline.
   • Check result/ for generated performance metrics.

---

🏗️ Project Structure Overview

1. CodeBase This is the "Clean Room." Only finalized code goes here.

• data/: Raw datasets (student-mat.csv, student-por.csv).
• src/: Source code, notebooks, and main scripts.
• result/: Model performance CSVs and outputs.

2. Literature (Collaborator Workspaces) This is the "Sandbox."

• Papers/: Shared collection of research papers we are studying (e.g., Evaluating the Explainers, Fairness in Student Prediction).
• Individual Workspaces:
  • 👤 Istiaq: Working on LSTM models, XAI methods (SHAP, LIME, DICE), and fairsynedu.
  • 👤 Nazme: Literature review and initial drafts.
  • 👤 Maria: Summaries of academic performance prediction papers.

3. Thesis_Paper Contains the academic output.

• LaTeX source files.
• High-resolution figures generated from the CodeBase.

---

🛠️ Key Technologies & Methods

• Models: LSTM, Bi-LSTM, Diffusion Models.
• Explainability (XAI): SHAP, LIME, DICE, CEM (Counterfactual Explanations).
• Fairness: Bias detection in educational data (fairsynedu).
• Data: Student Performance Data Set (Cortez and Silva, 2008).

---

🤝 Contributors

• Istiaq - Model Development & XAI Implementation
• Nazme - Literature Review & Documentation
• Maria - Research Analysis & Summarization

---

📝 License

This project is for academic research purposes.