Enhancing privacy while preserving the accuracy of collaborative filtering

Shlomo Berkovsky; Tsvi Kuflik; Francesco Ricci

Enhancing privacy while preserving the accuracy of collaborative filtering

Shlomo Berkovsky, Tsvi Kuflik, Francesco Ricci

Research output: Chapter in Book/Report/Conference proceeding › Conference proceeding contribution › peer-review

Abstract

Collaborative Filtering (CF) is considered a powerful technique for generating personalized recommendations. Centralized storage of user profiles in CF systems presents a privacy breach, since the profiles are available to other users. Recent works proposed enhancing the privacy of the CF by distributing the profiles between multiple repositories. This work investigates how a
decentralized distributed storage of user profiles combined with data perturbation techniques mitigates the privacy issues. Experiments, conducted on three datasets, show that relatively large parts of the profiles can be perturbed without hampering the accuracy of the CF. The experiments also allow conclusion to be drawn regarding the specific users and parts of the profiles that are valuable for generating accurate CF predictions.

Original language	English
Title of host publication	Proceedings of the ECAI 2006 Workshop on Recommender Systems
Pages	49-53
Number of pages	5
Publication status	Published - 2006
Externally published	Yes
Event	17th European Conference on Artificial Intelligence (ECAI 2006) - Riva del Garda, Italy Duration: 29 Aug 2006 → 1 Sept 2006

Conference

Conference	17th European Conference on Artificial Intelligence (ECAI 2006)
Country/Territory	Italy
City	Riva del Garda
Period	29/08/06 → 1/09/06

Cite this

@inproceedings{5bbffb5f08884384aff3905ef0d10f8a,

title = "Enhancing privacy while preserving the accuracy of collaborative filtering",

abstract = "Collaborative Filtering (CF) is considered a powerful technique for generating personalized recommendations. Centralized storage of user profiles in CF systems presents a privacy breach, since the profiles are available to other users. Recent works proposed enhancing the privacy of the CF by distributing the profiles between multiple repositories. This work investigates how adecentralized distributed storage of user profiles combined with data perturbation techniques mitigates the privacy issues. Experiments, conducted on three datasets, show that relatively large parts of the profiles can be perturbed without hampering the accuracy of the CF. The experiments also allow conclusion to be drawn regarding the specific users and parts of the profiles that are valuable for generating accurate CF predictions.",

author = "Shlomo Berkovsky and Tsvi Kuflik and Francesco Ricci",

year = "2006",

language = "English",

pages = "49--53",

booktitle = "Proceedings of the ECAI 2006 Workshop on Recommender Systems",

note = "17th European Conference on Artificial Intelligence (ECAI 2006) ; Conference date: 29-08-2006 Through 01-09-2006",

}

TY - GEN

T1 - Enhancing privacy while preserving the accuracy of collaborative filtering

AU - Berkovsky, Shlomo

AU - Kuflik, Tsvi

AU - Ricci, Francesco

PY - 2006

Y1 - 2006

N2 - Collaborative Filtering (CF) is considered a powerful technique for generating personalized recommendations. Centralized storage of user profiles in CF systems presents a privacy breach, since the profiles are available to other users. Recent works proposed enhancing the privacy of the CF by distributing the profiles between multiple repositories. This work investigates how adecentralized distributed storage of user profiles combined with data perturbation techniques mitigates the privacy issues. Experiments, conducted on three datasets, show that relatively large parts of the profiles can be perturbed without hampering the accuracy of the CF. The experiments also allow conclusion to be drawn regarding the specific users and parts of the profiles that are valuable for generating accurate CF predictions.

AB - Collaborative Filtering (CF) is considered a powerful technique for generating personalized recommendations. Centralized storage of user profiles in CF systems presents a privacy breach, since the profiles are available to other users. Recent works proposed enhancing the privacy of the CF by distributing the profiles between multiple repositories. This work investigates how adecentralized distributed storage of user profiles combined with data perturbation techniques mitigates the privacy issues. Experiments, conducted on three datasets, show that relatively large parts of the profiles can be perturbed without hampering the accuracy of the CF. The experiments also allow conclusion to be drawn regarding the specific users and parts of the profiles that are valuable for generating accurate CF predictions.

M3 - Conference proceeding contribution

SP - 49

EP - 53

BT - Proceedings of the ECAI 2006 Workshop on Recommender Systems

T2 - 17th European Conference on Artificial Intelligence (ECAI 2006)

Y2 - 29 August 2006 through 1 September 2006

ER -

Enhancing privacy while preserving the accuracy of collaborative filtering

Abstract

Conference

Fingerprint

Cite this