Fuzzy logic for decision extraction from product reviews

Abstract

Most of manufacturers and merchants sell their products via internet websites where customers buy them and write their reviews regarding the purchased product. These reviews carry a lot of information. Prospective customers are heavily interested in the experience of the predecessor customers where enormous number of reviews are offered by the website per product. In this study given a dataset of reviews we give a solution to estimate a total score of sentiment for the product reviewed. For this purpose fuzzy logic is used to process the subjective comments and to obtain a reliable average estimation for the decision from those individual reviews. The method is executed over different models to search the effects of the frequency of features on the outcome. © 2017 IEEE.