A machine learning approach for the identification of the deceptive reviews in the hospitality sector using unique attributes and sentiment orientation

The popularity of online reviews is causing a huge impact on consumers’ purchase intentions for goods and services. However, and hidden by the anonymity of the Internet, fraudsters can try to manipulate other consumers by posting fake reviews. Maintaining trust in online reviews require the development of automatic tools using machine learning approaches because of the huge volume of online opinions generated every day. This paper is focused on the hospitality sector and follows a content analysis approach based on a set of unique attributes and the sentiment orientation of reviews. The main contributions of the paper are i) a set of polarity-oriented unique attributes able to distinguish positive and negative deceptive and non-deceptive reviews and ii) the main topics associated to positive and negative deceptive and non-deceptive reviews. Findings reveal that positive and negative unique attributes lead to non-biased classifiers and that experience based reviews tend to be non-deceptive.