An improved privacy-preserving DWT-based collaborative filtering scheme

Bilge, Alper; Polat, Hüseyin

doi:10.1016/j.eswa.2011.09.094

Cited by 24 publications

(14 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Amount of traditional privacy preserving methods have been developed in CF recommender systems [27], including cryptographic [9,19], obfuscation [20,25], perturbation [3,4] and probabilistic methods [1]. Erkin et al [9] applied homomorphic encryption and secure multi-party computation in privacy preserving recommender systems, which allows users to jointly compute their data to receive recommendation without sharing the true data with other parties.…”

Section: Traditional Privacy Preserving Cf Recommendationmentioning

confidence: 99%

“…Among them, cryptographic methods [9,19] provide the most reliable security but the unnecessary computational cost cannot be ignored. Obfuscation methods [20,25] and Perturbation methods [3,4] introduce designed random noise into the original matrix to preserve customers' sensitive information; however the magnitude of noise is hard to calibrate in these two types of methods [7,27]. The naive probabilistic method [1] provides a similarity based weighted neighbour selection for the k neighbours.…”

mentioning

confidence: 99%

See 1 more Smart Citation

A Security-assured Accuracy-maximised Privacy Preserving Collaborative Filtering Recommendation Algorithm

Lü

Shen

2014

Proceedings of the 19th International Database Engineering &Amp; Applications Symposium on - IDEAS '15

View full text Add to dashboard Cite

The neighbourhood-based Collaborative Filtering is a widely used method in recommender systems. However, the risks of revealing customers' privacy during the process of filtering have attracted noticeable public concern recently. Specifically, kNN attack discloses the target user's sensitive information by creating k fake nearest neighbours by nonsensitive information. Among the current solutions against kNN attack, the probabilistic methods showed a powerful privacy preserving effect. However, the existing probabilistic methods neither guarantee enough prediction accuracy due to the global randomness, nor provide assured security enforcement against kNN attack. To overcome the problems of current probabilistic methods, we propose a novel approach, Partitioned Probabilistic Neighbour Selection, to ensure a required security guarantee while achieving the optimal prediction accuracy against kNN attack. In this paper, we define the sum of k neighbours' similarity as the accuracy metric α, the number of user partitions, across which we select the k neighbours, as the security metric β. Differing from the present methods that globally selected neighbours, our method selects neighbours from each group with exponential differential privacy to decrease the magnitude of noise. Theoretical and experimental analysis show that to achieve the same security guarantee against kNN attack, our approach ensures the optimal prediction accuracy.

show abstract

Section: Traditional Privacy Preserving Cf Recommendationmentioning

confidence: 99%

mentioning

confidence: 99%

A Security-assured Accuracy-maximised Privacy Preserving Collaborative Filtering Recommendation Algorithm

Lü

Shen

2014

Proceedings of the 19th International Database Engineering &Amp; Applications Symposium on - IDEAS '15

View full text Add to dashboard Cite

show abstract

“…Due to privacy measures, performance might become worse. To enhance online performance in PPCF schemes on RPTs, Bilge and Polat [14] propose dimensionality reduction-based PPCF scheme. The authors apply data reduction on perturbed data and provide predictions from reduced masked data.…”

Section: Introductionmentioning

confidence: 99%

Privacy-preserving normalized ratings-based weighted slope one predictor

Terzi¹,

Polat²

2016

Int. J. DNE

View full text Add to dashboard Cite

Weighted Slope One predictor is proposed as a model-based collaborative filtering algorithm based on user ratings. The predictor is able to efficiently provide accurate predictions. The scheme utilizes user's true ratings. In this paper, we propose to utilize normalized user ratings like z-scores for the weighted Slope One predictor. Also, in order to protect privacy, we propose a privacy-preserving weighted Slope One predictor based on z-scores using randomization. Moreover, we utilize masked deviations to show how it affects accuracy of the proposed scheme. We perform various real data-based experiments to evaluate the overall performance of the proposed method. Empirical outcomes show that the algorithm is able to provide accurate predictions.

show abstract

“…Among them, cryptographic methods [11,21] provide the most reliable security but the unnecessary computational cost cannot be ignored. Obfuscation methods [22,27] and Perturbation methods [3,4] introduce designed random noise into the original matrix to preserve customers' sensitive information; however the magnitude of noise is hard to calibrate in these two types of methods [9,29]. The probabilistic methods [1] provided a similarity based weighted neighbour selection of the k nearest neighbours.…”

Section: Introductionmentioning

confidence: 99%

An accuracy-assured privacy-preserving recommender system for internet commerce

Lü

Shen

2015

COMSIS J

View full text Add to dashboard Cite

Recommender systems, tool for predicting users' potential preferences by computing history data and users' interests, show an increasing importance in various Internet applications such as online shopping. As a well-known recommendation method, neighbourhood-based collaborative filtering has attracted considerable attention recently. The risk of revealing users' private information during the process of filtering has attracted noticeable research interests. Among the current solutions, the probabilistic techniques have shown a powerful privacy preserving effect. The existing methods deploying probabilistic methods are in three categories, one [19] adds differential privacy noises in the covariance matrix; one [1] introduces the randomisation in the neighbour selection process; the other [29] applies differential privacy in both the neighbour selection process and covariance matrix. When facing k Nearest Neighbour (kNN) attack, all the existing methods provide no data utility guarantee, for the introduction of global randomness. In this paper, to overcome the problem of recommendation accuracy loss, we propose a novel approach, Partitioned Probabilistic Neighbour Selection, to ensure a required prediction accuracy while maintaining high security against kNN attack. We define the sum of k neighbours' similarity as the accuracy metric α, the number of user partitions, across which we select the k neighbours, as the security metric β. We generalise the k Nearest Neighbour attack to βk Nearest Neighbours attack. Differing from the existing approach that selects neighbours across the entire candidate list randomly, our method selects neighbours from each exclusive partition of size k with a decreasing probability. Theoretical and experimental analysis show that to provide an accuracy-assured recommendation, our Partitioned Probabilistic Neighbour Selection method yields a better trade-off between the recommendation accuracy and system security.

show abstract

An improved privacy-preserving DWT-based collaborative filtering scheme

Cited by 24 publications

References 42 publications

A Security-assured Accuracy-maximised Privacy Preserving Collaborative Filtering Recommendation Algorithm

A Security-assured Accuracy-maximised Privacy Preserving Collaborative Filtering Recommendation Algorithm

Privacy-preserving normalized ratings-based weighted slope one predictor

An accuracy-assured privacy-preserving recommender system for internet commerce

Contact Info

Product

Resources

About