🛒 Building and Optimizing a Personalized Recommendation System for E-Commerce using Machine Learning

Hi! I'm Hari Gharti Magar, an MSc Data Science with Advanced Research student at the University of Hertfordshire, UK. This project was developed as part of my final dissertation, focusing on "Building and Optimizing a Personalized Recommendation System for E-Commerce Using Machine Learning".

---

📘 Project Summary

This project implements three recommendation system models to suggest products to users:

• Collaborative Filtering using SVD (Singular Value Decomposition)
• Content-Based Filtering using TF-IDF and Nearest Neighbors
• Hybrid Model combining both approaches

Additionally, I built a functional e-commerce website prototype to simulate real-time user interaction and generate purchase data for model training.

---

📦 Dataset

Dataset used:
E-Commerce Purchase History from Electronics Store
📎 Kaggle Link

• Format: CSV
• Records: 1.3 million+
• License: Free for educational and research use

---

📊 Results

Model	Accuracy	Diversity	Engagement
Collaborative Filter	RMSE ≈ 283.5	✅ High	✅ High
Content-Based Filter	Match Rate ≈ 55%	❌ Low	❌ Low
Hybrid Model	Match Rate ≈ 55%	❌ Low	❌ Low

---

🛠️ Technologies Used

• Languages: Python, PHP, HTML, CSS, JavaScript
• Libraries: Scikit-learn, Surprise, Pandas, NumPy, TF-IDF, Seaborn
• Tools: Jupyter Notebook (Google Colab), MySQL, XAMPP, Visual Studio Code

---

📁 Project Structure

📄 README.md → Project overview and documentation 📄 RecommendationSystem.ipynb → Jupyter Notebook for model development 📄 RecommendationSystem.ipynb - Colab.pdf → PDF version of the notebook (for quick view) 📄 Report_Recommendation_System.pdf → Final MSc dissertation report (PDF) 📄 recommendationsystem.py → Python script version of the notebook

🌐 E-Commerce Prototype

As part of this project, I also developed a working e-commerce website prototype that simulates product browsing, purchase activity, and real-time user behavior for data generation and testing.

🔗 View E-Commerce Prototype Repository

Tech Stack:

• Frontend: HTML, CSS, JavaScript
• Backend: PHP & MySQL
• Local Testing: XAMPP

---

📄 Report

📥 Download Full Project Report

The report includes research background, methodology, machine learning implementation, evaluation metrics, and screenshots of the working web system.

---

🙋 About Me

• MSc Data Science with Advanced Research (2023–2025)
• University of Hertfordshire, UK
• Passionate about recommendation systems, applied machine learning, and solving real-world problems with data.

📧 magarharee54@gmail.com
🔗 LinkedIn 💻 GitHub

---

Thanks for visiting my project!