Beware of the Fakes - Overview of Fake Detection Methods for Online Product Reviews

Abstract

Online reviews about products and services, such as reviews in stores, are a valuable source of information for customers. Unfortunately, reviews are contaminated by fake reviews, which may lead to wrong conclusions when including them in the analyses of user feedback. As these fake reviews are not marked as advertisement, they might lead to wrong conclusions for customers. If customers are trusting fake reviews their user experience is significantly lowered as soon as they find out that they were betrayed. Therefore, online stores and social media platforms have to take countermeasures against fake reviews. Thus, we performed a systematic literature review to create an overview of the available methods to detect fake reviews and relate the methods to their necessarily required data. This will enable us to identify fake reviews within different data sources easier in order to improve the reliability of the used customer feedback. We have analyzed 141 methods for fake detection. As the reporting quality of a substantial part lacked understandability in terms of method description and evaluation details, we have provided recommendations for method and evaluation descriptions for future method proposals. In addition, we have performed an assessment in terms of detection effectiveness and quality of those methods.