Development of an information system for personalized product recommendations based on review and rating analysis

Abstract

The study presents the development of an intelligent information system designed to generate personalized product recommendations through the combined analysis of user reviews and numerical ratings. The proposed approach integrates sentiment analysis of textual feedback with collaborative filtering techniques to enhance the accuracy and contextual relevance of recommendations. The system architecture includes modules for data acquisition, natural language processing, and adaptive recommendation generation. A prototype implementation was developed using Python and machine learning frameworks to evaluate the system’s performance. Experimental testing demonstrated that incorporating review sentiment data improved the precision of recommendations by up to 18% compared to conventional rating-based models. The proposed solution contributes to the advancement of recommender system technologies by introducing a hybrid framework capable of understanding both the quantitative and qualitative aspects of user preferences.