Sentiment Analysis and Innovative Recommender System: Enhancing Goodreads Book Discovery Using Hybrid Collaborative and Content Based Filtering

This study addresses the challenge of finding suitable books in the digital age of education, where information overload makes manual book selection difficult. It aims to analyze reviewer sentiment on Goodreads and develop a recommendation algorithm based on readers’ preferences. The study employs sentiment analysis and compares three recommender algorithms: Content-Based, Collaborative filtering, and Hybrid filtering. Goodreads data is collected using a web scraper, and the results indicate Hybrid filtering as the most effective model, outperforming others in metrics like RMSE, MSE, precision, and recall. Further optimization with the Apriori model can enhance Hybrid filtering’s accuracy and recommendation breadth, reducing system errors.