A Comparative Evaluation of Profile
Injection Attacks
Anjani Kumar Verma and Veer Sain Dixit
Abstract In recent years, the research on shilling attacks has been greatly improved.
However, some serious problem in hand such as attack model dependency and high
computational cost. Such recommender system also provides an impressive way
to overcome information overload problem. In order to preserve the trust of rec-
ommender system, it is required to identify and remove the fictitious profiles from
the system. Here, we have used machine learning classifiers to detect the attacker’s
profiles. A new model is proposed that outperforms in most of the cases.
Keywords Recommender systems · Collaborative filtering · Shilling attacks
MAE · RMSE
1 Introduction
In recent years, the effective profile injection or shilling attacks are more emphasized
toward the insertion of bogus user profiles into the system database in order to manip-
ulate the recommendation output, which is actually used to promote or demote the
predicted ratings for a particular product. In many e-commerce websites, the recom-
mender systems are widely deployed to provide user-purchasing suggestion. With
rapid change in technology, most recommender systems adopted collaborative filter-
ing. However, with the open nature of collaborative filtering recommender systems, it
suffers from significant vulnerabilities from being attacked by malicious raters, who
inject profiles consisting of biased ratings. Hence, they may have an effective impact
on produced predictions. The basic function of a recommender system is to sense
the consumer’s feedback (which can be present in many forms like implicit, explicit,
etc.) and understand user’s interests to benefit consumers and e-business owners.
A. K. Verma (B)
Department of Computer Science, University of Delhi, New Delhi, India
e-mail: anjaniverma29@gmail.com
V. S. Dixit
Department of Computer Science, ARSD College, University of Delhi, New Delhi, India
e-mail: veersaindixit@rediffmail.com
© Springer Nature Singapore Pte Ltd. 2019
M. L. Kolhe et al. (eds.), Advances in Data and Information Sciences, Lecture Notes
in Networks and Systems 39, https://doi.org/10.1007/978-981-13-0277-0_4
43