Skip to content

Leo-bsb/book-recommendation-system

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 

Repository files navigation

Intelligent Book Recommendation System

Python PyTorch License GitHub Repo Size Last Commit Open in Colab


A hybrid book recommendation system that leverages deep learning, combining collaborative filtering and content-based approaches to deliver highly personalized and diverse book suggestions.

🔗 Try it Online (Hugging Face Space) 🔗 Open in Google Colab

Note: The interface and notebook content/documentation are in Portuguese (PT-BR).


🎯 Overview

This system harnesses the power of deep neural networks to analyze user behavior and book features, providing accurate, scalable, and diverse recommendations. Using PyTorch, it implements a neural collaborative filtering framework enhanced by content-based insights for optimal user experience.


📊 Dataset

GoodBooks-10k — one of the most comprehensive publicly available book rating datasets:

  • Over 6 million ratings
  • 10,000 distinct books
  • 53,424 active users
  • Ratings on a 1-5 star scale

🛠️ Tech Stack

  • Programming Language: Python 3.12
  • Deep Learning: PyTorch
  • Data Processing: Pandas, Polars, NumPy
  • Machine Learning: Scikit-learn
  • Deployment: Gradio for interactive demo
  • Cloud: Google Colab for easy experimentation

✨ Key Features

  • Hybrid Recommendation: Fuses collaborative filtering and content-based techniques
  • Deep Neural Network: Learns complex user-item interaction patterns
  • Personalization: Delivers tailored suggestions for each user
  • Scalability: Efficient handling of large-scale datasets and users
  • Interactive Demo: Seamless web interface for exploration

🚀 Getting Started

  1. Open the Google Colab notebook
  2. Run all cells sequentially
  3. Experiment with parameters and observe recommendation changes
  4. Deploy the Gradio demo or integrate into your own system

📈 Model Architecture

  • Embedding Layers: Learn latent factors for users and books
  • Neural Collaborative Filtering: Multi-layer perceptron capturing nonlinear user-item relationships
  • Matrix Factorization: Foundation for interaction modeling
  • Hybrid Scoring: Combines collaborative and content-based scores for balanced recommendations

📖 Citation

If you use this project in your research or work, please cite accordingly.

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors