Ranking products with online reviews: A novel method based on hesitant fuzzy set and sentiment word framework

A novel method to rank products through online reviews by view different sentiment scores of one feature as the different membership degrees, and extends TODIM method, which concerns decision make’s psychological behavior, to deal with criteria interactions in the process of MADM.

Abstract

Recently, sentiment analysis (SA) and multi-attribute decision making (MADA) have been extensively studied respectively, which aims to help decision makers make informed decisions. However, rather less attention has been paid to the field of combining SA and MADA. Therefore, in this paper, we propose a novel method to rank products through online reviews. To begin with, it is a novel idea to view different sentiment scores of one feature as the different membership degrees. Further, we propose the fuzzy sentiment word framework and corresponding computation rules to calculate the sentiment score of each feature in each review, which later can be used to obtain the overall performance of each feature concerning different products based on hesitant fuzzy set (HFS). Next, the attention degree of each feature is considered in the process of calculating weight of different features. In addition, based on 2-addiitive fuzzy measure and Choquet integral, we extend TODIM (an acronym in Portuguese of interactive and multi-criteria decision making) method, which concerns decision make’s psychological behavior, to deal with criteria interactions (positive, mutual independent and negative) in the process of MADM. Furthermore, we use a case study to demonstrate the efficiency and applicability of the proposed method.