https://github.com/pngo1997/hyperlink-analysis-content-based-recommendation-system

Hyperlink Analysis using PageRank and HITS algorithms and simple Content-Based Recommendation System.
https://github.com/pngo1997/hyperlink-analysis-content-based-recommendation-system

content-based-recommendation hits-algorithm knn link-analysis pagerank python

Last synced: 14 days ago
JSON representation

Hyperlink Analysis using PageRank and HITS algorithms and simple Content-Based Recommendation System.

• Host: GitHub
• URL: https://github.com/pngo1997/hyperlink-analysis-content-based-recommendation-system
• Owner: pngo1997
• Created: 2025-01-31T19:40:47.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2025-01-31T19:45:38.000Z (over 1 year ago)
• Last Synced: 2025-08-02T06:36:53.998Z (10 months ago)
• Topics: content-based-recommendation, hits-algorithm, knn, link-analysis, pagerank, python
• Language: Jupyter Notebook
• Homepage:
• Size: 230 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# 🔗 Hyperlink Analysis & Content-Based Recommendation System  

## 📜 Overview  

This project is divided into two main components:  

1. **Hyperlink Analysis** using **PageRank and HITS algorithms**.  

2. **Content-Based Recommendation System** using **K-Nearest Neighbors (KNN)** for song recommendations based on Spotify metadata.  

📌 **Datasets Used**:  

- **Hyperlink Analysis**: Simulated Web graph structure.  

- **Content-Based Recommendation**:  

  - `Spotify_Metadata.csv` – Song metadata.  

  - `Spotify_Numeric.csv` – Song features (numeric vectors).  

  - `Spotify_Users.txt` – User playlists.  

📌 **Programming Language**: `Python 3`  

📌 **Libraries Used**: `NumPy`, `pandas`, `networkx`, `scikit-learn`, `math`, `csv`  

## 🔗 1️⃣ Hyperlink Analysis  

### **1.1 PageRank Algorithm**  

- **Input**: Web graph structure: A → {B, C} B → {C} C → {A}

- **Computation**:  

- **Damping Factor (α) = 0.15**  

- **Three Iterations**  

- PageRank scores computed **before & after normalization** at each step.  

### **1.2 HITS Algorithm (Hubs & Authorities)**  

- **Input**: Web graph structure (including Page D).  

- **Computation**

- **Three Iterations**  

- Hub & Authority scores computed **before & after normalization** at each step.  

📌 **Output**:  

- PageRank & HITS scores printed at each iteration.  

## 🎵 2️⃣ Content-Based Recommendation System  

### **2.1 K-Nearest Neighbors (KNN) for Song Similarity**  

- Function: `knn(item, Data, K)`  

- **Input**:  

  - `item` – Feature vector of a song.  

  - `Data` – Song feature matrix (`Spotify_Numeric.csv`).  

  - `K` – Number of nearest neighbors.  

- **Computation**:  

  - Compute **Cosine Similarity** between `item` and all songs in `Data`.  

  - Return **K most similar songs** with metadata (`Spotify_Metadata.csv`).  

📌 **Demonstration**:  

- Run `knn()` for the given song vector: item = [0.8, 0.1, 0.7, 0.0, 0.5, 0.1, 0.5, 0.2, 0.0, 1.0, 0.1, 0.2, 0.7]

- Print **top 10 similar songs**.  

### **2.2 User-Based Song Recommendations**  

- Function: `recommend(userID, Data, K)`  

- **Input**:  

  - `userID` – User index (0-based).  

  - `Data` – Song feature matrix.  

  - `K` – Number of recommendations.  

- **Computation**:  

  - Compute **user profile** as the **centroid vector** of liked songs.  

  - Call `knn()` to get **top K song recommendations**.  

  - **Exclude previously listened songs** from final recommendations.  

📌 **Demonstration**:  

- Run `recommend()` for:  

- **User 3** (4th row in `Spotify_Users.txt`)  

- **User 19** (20th row in `Spotify_Users.txt`)  

- Print **top 20 recommendations** for each user.