Using Machine Learning Algorithms to Alleviate the Shilling Attack in a Recommendation System

Abstract

Abstract: Every e-commerce site must have a referral system. Shilling attacks, One of the biggest issues with recommendation systems is those that involve the creation of fake profiles in the system and biased ratings of things. They reduce the accuracy and improve the performance of the recommender system when making suggestions to users. The goal of attackers is to change the order of materials or objects that match their interests. Shilling attacks threaten the reliability of RS. Therefore, to maintain their validity and fairness, recommender systems must be able to detect shilling attacks. So far, suitable algorithms and methods have been presented for the detection of shilling attacks. Some of these approaches, though, either use low-order interactions or higher-order interactions, or they examine the rating matrix from a single point of view. This study aimed to propose a mechanism using users' rating matrix, rating time, and social network analysis output of users' profiles by Gaussian-Rough neural network to simultaneously use low-order and high-order interactions to detect shilling attacks. Finally, several experiments were conducted with three models: CNN, DNN, and RNN compared with the Proposed Model. The results indicated that the proposed method is more effective than the comparison methods regarding attack detection and overall detection, which proves the effectiveness of the proposed model.