Book Recommender System Using Content-Based Filtering for PNJ Press Website
Dewi Yanti Liliana, Rizki Elisa Nalawati, Ratna Widya Iswara, Alifiah Putri Aisyah, Hana Octavia Trinida Malo · 2024
The vast collections in book publishers and bookstores make it challenging for users to find books on specific topics without knowing the exact titles. An information system with a book recommendation feature can significantly enhance the efficiency of a book search. Recommender systems, which predict or suggest items of interest based on user interactions and queries, are widely used across various platforms to personalize user experiences. This study focuses on developing a Content-Based Filtering recommender system for PNJ Press, a book publisher of the academic community of Jakarta State Polytechnic. The system compares user search queries with the content of book titles and abstracts to provide relevant recommendations. Previous research has demonstrated the effectiveness of Content-Based Filtering in various domains, highlighting its potential for improving book search efficiency on the PNJ Press website. The system achieved a precision of 91.84% and a recall of 97.83%, resulting in an overall accuracy rate of 90%. Additionally, the study found that the combined attributes of book titles and abstracts significantly influence recommendation results.