Robust recommendation algorithm using an iterative group-based reputation

Abstract

Collaborative filtering recommender systems have been widely applied in e-commerce and have brought huge economic benefits. Malicious users can deliberately inject attack profiles into a recommender system to affect the recommendation result. Detecting and defending against shilling attacks in recommender systems is very important. The existing algorithms, which are based on the latent factor model, have poor robustness and low recommendation accuracy under shilling attacks. Thus, we proposed a robust collaborative recommendation algorithm that is based on users' reputations, which are determined based on their grouping behaviours. We group users according to their rating histories and iteratively compute each user's reputation. This algorithm has two advantages: It has high recommendation accuracy and stronger robustness and does not destroy the user distribution of the original dataset. Experiments on the real datasets MovieLens 100K and MovieLens 1M show that the proposed robust algorithm outperforms previous robust algorithms in terms of both accuracy and robustness.