Spam Detection Approach for Cloud Service Reviews Based on Probabilistic Ontology

Online reviews provide a vision on the strengths and weakness of products/services, influencing potential customers’ purchasing decisions. The fact that anybody can leave a review provides the opportunity for spammers to write spam reviews about products and services for different intents. To counter this problem, a number of approaches for detecting spam reviews have been proposed. However, to date, most of these approaches depend on rich/complete information about items/reviewers, which is not the case of Social Media Platforms (SMPs). In this paper, we consider well known spam features taken from the literature to them we add two new ones: the user profile authenticity to allow the detection of spam review from any SMP and opinion deviation to verify the opinion truthfulness. To define a common model for different SMPs and to cope with the incompleteness of information and uncertainty in spam judgment, we propose a Review Spam Probabilistic Ontology (RSPO) based approach. Probabilistic Ontology is defined using Probabilistic Web Ontology Language (PR-OWL) and the probability distributions of the review spamicity is defined automatically using a learning approach. The herein reported experimental results proved the effectiveness and the performance of the approach.