From existing trends to future trends in privacy‐preserving collaborative filtering

Abstract

The information overload problem, also known as infobesity, forces online vendors to utilize collaborative filtering algorithms. Although various recommendation methods are widely used by many electronic commerce sites, they still have substantial problems, including but not limited to privacy, accuracy, online performance, scalability, cold start, coverage, grey sheep, robustness, being subject to shilling attacks, diversity, data sparsity, and synonymy. Privacy-preserving collaborative filtering methods have been proposed to handle the privacy problem. Due to the increasing popularity of privacy protection and recommendation estimation over the Internet, prediction schemes with privacy are still receiving increasing attention. Because research trends might change over time, it is critical for researchers to observe future trends. In this study, we determine the existing trends in the privacypreserving collaborative filtering field by examining the related papers published mainly in the last few years. Comprehensive examinations of the most up-to-date related studies are described. By scrutinizing the contemporary inclinations, we present the most promising possible research trends in the near future. Our proposals can help interested researchers direct their research toward better outcomes and might open new ways to enrich privacy-preserving collaborative filtering studies.

Privacy-preserving collaborative recommendations based on random perturbations

“…Privacy is an important issue for users of such systems Ozturk and Polat, 2015). In the context of ratings, privacy concerns make users unwilling to submit ratings, thus leading to sparsely populated relevant datasets which in turn can lead to lower degrees of similarity and eventually to poor recommendations.…”

Collaborative Filtering Under a Sybil Attack: Similarity Metrics do Matter!

“…Our results go beyond the partial analysis of [17], highlight the tradeoff between privacy and recommendation quality, and provide a solution to it. b) Privacy Protection in Recommenders: Protecting collaborative filtering from the above attacks constitutes a promising research direction [32]. The first attempts to provide privacy-preserving recommenders focused on decentralized solutions based on homomorphic encryption [13], anonymization [11], or profile obfuscation [6].…”

Privacy-preserving normalized ratings-based weighted slope one predictor

“…The most important problems can be considered as accuracy, efficiency, and privacy. It is very important for CF schemes to provide accurate predictions efficiently while preserving privacy [4]. Due to increasing number of privacy risks posed by CF algorithms [4], users might hesitate to share their data with CF systems or refuse to provide data at all.…”

“…Privacy-preserving collaborative filtering (PPCF) schemes have been proposed to achieve privacy while providing accurate recommendations [2,4]. Canny [5,6] has initiated the collaborative filtering with privacy.…”

“…Canny [5,6] has initiated the collaborative filtering with privacy. Since then various studies on PPCF have been proposed in the literature [2,4]. One of the most common privacy protection methods is known as randomization.…”