An Unsupervised Approach for Reputation Generation

Nowadays, watching a movie, buying a product, making hotel reservations and other e-commerce trades are strung to consulting other peoples reviews and recommendations on the target entity. Indeed, Amazon, IMDB (Internet Movie Database) as well as several websites provide a convenient platform where users share freely their opinions and their subjective attitudes towards the target entity with no restrictions. However, those opinions are too much to be examined one by one, this is why a general reputation value makes the task of choosing the right product much easier. In this paper, we propose a reputation generation approach based on opinion clustering and semantic analysis. In our approach, opinions are grouped into a number of clusters that contain opinions with the same attitude or preference. By aggregating the ratings attached to the clusters, we generate the reputation of an entity. Experimental results demonstrate the effectiveness of the proposed approach in generating reputation value. (C) 2019 The Authors. Published by Elsevier B.V.