Book Recommendation System using Machine Learning

Abstract

The increasing volume of online texts made it imperative to develop intelligent systems that can help readers discover books of their choice. This article presents a Book Recommender System employing popularity-based and collaborative filtering algorithms to enhance personalized recommendations. The system employs datasets comprised of user ratings, book metadata, and user profiles to develop two distinct models. The popularity-based approach chooses top-rated books according to the number of ratings and mean rating, giving popular items to new users. The collaborative filtering approach employs user behavior to recommend books by discovering similar readers and their interests, giving a more personalized experience. Data preprocessing involved missing value handling, duplicate elimination, and removing users with large rating histories for improving recommendation accuracy. The models' performance was measured in terms of diversity and relevance of the suggested books. The popularity-based model effectively indicates best-selling books, and the collaborative filtering model indicated improved personalization using user interactions. This study emphasizes the significance of combining different recommendation methods to cater to diverse user needs and paves the way for further research in hybrid recommendation systems.