Rating prediction based on combination of review mining and user preference analysis

Review websites allow users to share their reviews of products or businesses, give ratings to products or businesses, and interact with other users. Due to the rapid growth of online review data, users face the problem of information overload. To resolve this problem, many researches have proposed various recommendation methods based on the analysis of users' ratings. Besides user ratings, the review websites contain unstructured textual data and information of different aspects which has different impact and importance to both users and businesses. It may lead to inaccurate rating predictions, as it is difficult to know users' preferring aspects and their corresponding importance by analyzing users' rating data. To resolve the above-stated problems, this research proposed aspect-based rating prediction methods, i.e., ARPM and ARPM-Social, which integrate aspect detection and sentiment analysis to generate user preference and business performance, combined with the results of social behavior analysis to predict the ratings of the businesses that users will be interested in the future. Based on the experimental results, the proposed methods perform better than other traditional rating-based prediction methods. They can effectively analyze various aspects and sentiments from the reviews of users and businesses, as well as improve the accuracy of rating predictions. (C) 2021 Elsevier Ltd. All rights reserved.

Automated summary

Review websites allow users to share their reviews and ratings, and interact with other users; to address the problem of information overload, various recommendation methods based on users' ratings analysis have been proposed; this research introduced aspect-based rating prediction methods, ARPM and ARPM-Social, which outperform traditional methods.