This project implements a book recommendation system using collaborative filtering with k-nearest neighbors (k-NN). The system predicts user ratings for books and provides recommendations based on user preferences. The Surprise library is used to build and evaluate the k-NN model. Python functions are included to deliver personalized book recommendations and rating predictions. The dataset can be obtained from here.
- Books Dataset: Contains information about 271,379 books, including attributes like ISBN, title, author, publication year, and publisher.
- Book Ratings Dataset: Includes 1,149,780 ratings given by users to various books. Each rating can be explicit (user-provided) or implicit.
- Users Dataset: Contains demographic information for 278,858 users, including user IDs, location, and age.
- Utilizes user-based collaborative filtering to recommend books based on user similarity.
- Implements the k-nearest neighbors algorithm to predict ratings and generate recommendations.
- Excludes books and users with insufficient interaction to enhance recommendation quality.
- Python: Version 3.6 or higher
- Jupyter Notebook: If using Google Colab or any notebook environment
- Required Libraries:
pandasnumpymatplotlibseabornscikit-surprise