Personalized Reviews Based on Aspect Analysis and Polarity

Before e-commerce site users make a decision to purchase a product, first of all, they read the reviews on the sites. Based on the votes which website owners receive from users that read reviews, the site owners choose the best reviews. Nevertheless, it has recently been found that redundant information may be contained in such reviews. Thus, it has been recommended that review selection is made based on coverage (i.e., the number of entity aspects) enveloped by the reviewers. This paper focuses on additional improving the topmost review group through the use of customization criteria. This driven by the relevance of product aspects may differ among users, and users prefer to pay attention to the aspects that are more relevant to them. The main aim of this work is to give priority to the individual preferences and sentiments of users in terms of reviews. This is achieved through the selection of a personalized top review set by Personalized Aspect Analysis Model (PAAM), which covers reviews about aspects that are of more relevance to the user. The effectiveness of the proposed approach on computing PAAM with high-level coverage, top quality, and relevance of the subjects that are of importance to the customer are demonstrated through the experimental evaluation.