With upwards of two million books being published each year, recommending new books to read has become quite an overwhelming task. Lists of bestsellers only scratch the surface, and rarely, if ever, personalized for individual readers.

For our senior comprehensive project, we created a user-specific book rating prediction system - a key component of a book recommendation system that’s ready to rise to the challenge. We implemented and evaluated a combination of the leading collaborative and content-based filtering techniques, and combined them into an effective hybrid system. Our final report and implementations can be downloaded from Download.