Representing sentiment analysis results of online reviews using interval type-2 fuzzy numbers and its application to product ranking

Online reviews are used as a data source to make a variety of management decisions. An important precondition when using online reviews for decision analysis is knowing how to represent the sentiment analysis results of a large volume of online reviews. Although several approaches have been proposed for this, none of these consider the limited accuracy rates of the sentiment analysis results, which can impact the quality of the decision analysis results. To this end, we propose a new approach for representing the sentiment analysis results using interval type-2 fuzzy numbers that considers the accuracy rates. In the proposed approach, the sentiment analysis results with a 100% accuracy rate are converted into a triangular fuzzy number, and those with limited accuracy rate are regarded as the uncertainty of the membership function. Then, the results with the limited accuracy rate are converted into an interval type-2 fuzzy number. We discuss the related theoretical analysis to illustrate the validity of the proposed approach. Following this, a method for product ranking based on online reviews is described. In the proposed method, the sentiment analysis results of online reviews are represented using interval type-2 fuzzy numbers. Finally, a case study is presented to illustrate the use of the proposed method. (C) 2019 Elsevier Inc. All rights reserved.