Privacy preference inference via collaborative filtering

Taraneh Khazaei, Lu Xiao, Robert E. Mercer, Atif Khan

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Scopus citations


Studies of online social behaviour indicate that users often fail to specify privacy settings that match their privacy behaviour. This issue has caused a dilemma whether to use publicly available data for targeted advertisements and personalization. As a possible approach to manage this dilemma, we propose a collaborative filtering method that exploits homophily to build a probabilistic model. Such a model can indicate the likelihood that a given public profile is meant to be private. Here, we provide the results of an analysis of a set of observable variables to be used in a neighbourhood-based manner. In addition, we establish a social graph augmented with privacy information. Users in the graph are then transformed into a set of latent features, uncovering informative factors to infer privacy preferences.

Original languageEnglish (US)
Title of host publicationProceedings of the 10th International Conference on Web and Social Media, ICWSM 2016
PublisherAAAI Press
Number of pages4
ISBN (Electronic)9781577357582
StatePublished - 2016
Externally publishedYes
Event10th International Conference on Web and Social Media, ICWSM 2016 - Cologne, Germany
Duration: May 17 2016May 20 2016


Other10th International Conference on Web and Social Media, ICWSM 2016

ASJC Scopus subject areas

  • Computer Networks and Communications

Fingerprint Dive into the research topics of 'Privacy preference inference via collaborative filtering'. Together they form a unique fingerprint.

Cite this