TY - CHAP
T1 - Privacy in Recommender Systems
AU - Jeckmans, Arjan
AU - Beye, Michael
AU - Erkin, Zekeriya
AU - Hartel, Pieter H.
AU - Lagendijk, Reginald L.
AU - Tang, Qiang
PY - 2013/1
Y1 - 2013/1
N2 - In many online applications, the range of content that is offered to users is so wide that a need for automated recommender systems arises. Such systems can provide a personalized selection of relevant items to users. In practice, this can help people find entertaining movies, boost sales through targeted advertisements, or help social network users meet new friends.
To generate accurate personalized recommendations, recommender systems rely on detailed personal data on the preferences of users. Examples are ratings, consumption histories, and personal profiles. Recommender systems are useful, however the privacy risks associated to gathering and processing personal data are often underestimated or ignored. Many users are not sufficiently aware if and how much of their data is collected, if such data is sold to third parties, or how securely it is stored and for how long.
This chapter aims to provide insight into privacy in recommender systems. First, we discuss different types of existing recommender systems. Second, we give an overview of the data that is used in recommender systems. Third, we examine the associated risks to data privacy. Fourth, relevant research areas for privacy-protection techniques and their applicability to recommender systems are discussed. Finally, we conclude with a discussion on applying and combining different privacy-protection techniques in real-world settings, making clear mappings to reflect typical relations between recommender system types, information types, particular privacy risks, and privacy-protection techniques.
AB - In many online applications, the range of content that is offered to users is so wide that a need for automated recommender systems arises. Such systems can provide a personalized selection of relevant items to users. In practice, this can help people find entertaining movies, boost sales through targeted advertisements, or help social network users meet new friends.
To generate accurate personalized recommendations, recommender systems rely on detailed personal data on the preferences of users. Examples are ratings, consumption histories, and personal profiles. Recommender systems are useful, however the privacy risks associated to gathering and processing personal data are often underestimated or ignored. Many users are not sufficiently aware if and how much of their data is collected, if such data is sold to third parties, or how securely it is stored and for how long.
This chapter aims to provide insight into privacy in recommender systems. First, we discuss different types of existing recommender systems. Second, we give an overview of the data that is used in recommender systems. Third, we examine the associated risks to data privacy. Fourth, relevant research areas for privacy-protection techniques and their applicability to recommender systems are discussed. Finally, we conclude with a discussion on applying and combining different privacy-protection techniques in real-world settings, making clear mappings to reflect typical relations between recommender system types, information types, particular privacy risks, and privacy-protection techniques.
KW - Privacy-protection techniques
KW - SCS-Cybersecurity
KW - Privacy
KW - Recommender systems
U2 - 10.1007/978-1-4471-4555-4_12
DO - 10.1007/978-1-4471-4555-4_12
M3 - Chapter
SN - 978-1-4471-4554-7
T3 - Computer Communications and Networks
SP - 263
EP - 281
BT - Social Media Retrieval
A2 - Ramzan, Naeem
A2 - van Zwol, Roelof
A2 - Lee, Jong-Seok
A2 - Clüver, Kai
A2 - Hua, Xian-Sheng
PB - Springer
CY - London
ER -