Unsupervised Shilling Attack Detection Model Based on Rated Item Correlation Analysis

Abstract

Collaborative filtering is one of the most effective methods to deal with the information overload problem by providing a suitable recommendation to users. Recent research indicates that collaborative filtering is vulnerable to shilling attack which aims at manipulating the recommendation result through injecting malicious profiles. Our previous study suggests that applying the rated item correlation to supervised learning increases the accuracy of shilling attack detection. However, label information collection is one drawback of the supervised learning. In this study, an unsupervised detection based on the rated item correlation analysis is devised. The influence of the parameters on the detection accuracy is also discussed. The experimental results demonstrate that our proposed unsupervised detection model achieve a satisfying performance in shilling detection in MovieLens 100K dataset.

Lab Members

Patrick Chan

Associate Professor, Vice Dean

Patrick Chan works on machine learning, deep learning, image processing, adversarial learning, and secure machine learning.