Detection of spam reviews using hybrid grey wolf optimizer clustering method

Currently, online reviews play an essential role in the decision-making of customers. Various online websites such as Amazon, Yelp, Google Plus, BookMyShow, Facebook, Twitter, etc., allow its users to generate huge bulk of data. The data is generated in the form of feedback/reviews, comments, or tweets. This data is helpful for organizations to improve the quality of their products. Due to dependency on these online reviews, spam reviews are generated pretentiously by some organizations and people concerning promotion or demotion of the prominence of any product, organization, or person. Thus, identifying spam or non-spam review by the naked eye is nearly impossible. Classifying the reviews manually is also highly speculative. Hence, to overcome this issue, a hybrid Grey Wolf Optimizer (GWOK) based clustering method is proposed in this paper to identify spam reviews. In the proposed GWOK, the k-Means algorithm is used for initialization of the initial population for the basic GWO algorithm, and then the GWO algorithm is used for finding the optimal Cluster Heads. To prove that the proposed strategy is effective, three spam datasets, namely Synthetic Spam Reviews, Movie Reviews, and Yelp Hotel & Restaurant Reviews, have been used in our work. The reported results are compared with the existing state-of-art metaheuristic clustering methods like a genetic algorithm (GA), differential evolution (DE), particle swarm optimization (PSO), cuckoo search (CS), and k-Means. The results obtained by experimental and statistical analysis legitimize that the proposed GWOK algorithm surpasses contemporary techniques.