Securing Collaborative Filtering Against Malicious Attacks Through Anomaly Detection

Download: PDF.

“Securing Collaborative Filtering Against Malicious Attacks Through Anomaly Detection” by Runa Bhaumik, Chad Williams, Bamshad Mobasher, and Robin Burke. In Proceedings of the 4th Workshop on Intelligent Techniques for Web Personalization (ITWP'06), (Held at AAAI 2006, Boston, Massachusetts), July 2006.

Abstract

Collaborative filtering recommenders are highly vulnerable to malicious attacks designed to affect predicted ratings. Previous work related to detecting such attacks has focused on detecting profiles. Approaches based on profile classification to a large extent depend on profiles conforming to known attack models. In this paper we examine approaches for detecting suspicious rating trends based on statistical anomaly detection. We empirically show these techniques can be highly successful at detecting items under attack and time intervals when an attack occurred. In addition we explore the effects of rating distribution on detection performance and show that this varies based on distribution characteristics when these techniques are used.

Keywords: profile injection attacks, attack detection, collaborative filtering, recommender systems, time series, anomaly detection

Download: PDF.

BibTeX entry:

@inproceedings{BWMB06,
   author = {Runa Bhaumik and Chad Williams and Bamshad Mobasher and Robin
	Burke},
   title = {Securing Collaborative Filtering Against Malicious Attacks
	Through Anomaly Detection},
   booktitle = {Proceedings of the 4th Workshop on Intelligent Techniques
	for Web Personalization (ITWP'06)},
   address = {Held at AAAI 2006, Boston, Massachusetts},
   month = jul,
   year = {2006},
   url = {http://www.aaai.org/Press/Reports/Workshops/ws-06-10.php}
}