Robust Collaborative Recommendation by Least Trimmed Squares Matrix Factorization

Cheng, Zunping; Hurley, Neil

doi:10.1109/ictai.2010.90

Cited by 25 publications

(18 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In order to evaluate the performance of the proposed method (MTD-RMF), we carry out a series of contrast experiments with the existing robust recommendation algorithms: RMF [7], LTSMF [8] and VarSelect SVD [10]. The results of contrast experiments are shown in Table 1 and Table 2.…”

Section: Experimental Results and Analysismentioning

confidence: 99%

“…In order to improve the attack-resistant ability of collaborative recommender systems, researchers have proposed many shilling attack detection algorithms [4][5][6] and robust recommendation algorithms [7][8][9][10]. Mehta et al proposed shilling attack detection algorithm using principal component analysis (PCA) of the users' profiles [4].…”

Section: Introductionmentioning

confidence: 99%

“…Mehta et al proposed a robust recommendation algorithm (RMF) based on M-estimators [7]. The reference [8] proposed a least trimmed squares estimator based matrix factorization (LTSMF) algorithm, which shows better robustness compared with RMF. This kind of robust recommendation algorithm discards larger residuals ratings to resist malicious attacks, but it cause the loss of ratings information for genuine users.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Robust Recommendation Method Based on Shilling Attack Detection and Matrix Factorization Model

Hu¹,

Li²,

Zhang³

2018

dtcse

View full text Add to dashboard Cite

Abstract. The existing robust collaborative recommendation algorithms have low robustness against PIA and AoP attacks. Aiming at the problem, we propose a robust recommendation method based on shilling attack detection and matrix factorization model. Firstly, the type of shilling attack is identified based on statistical characteristics of attack profiles. Secondly, we devise corresponding unsupervised detection algorithms for standard attack, AoP and PIA, and the suspicious users and items are flagged. Finally, we devise a robust recommendation algorithm by combining the proposed shilling attack detection algorithm with matrix factorization model, and conduct experiments on the MovieLens dataset to demonstrate its effectiveness. Experimental results show that the proposed method exhibits good recommendation precision and excellent robustness for shilling attacks of multiple types.

show abstract

Section: Experimental Results and Analysismentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Robust Recommendation Method Based on Shilling Attack Detection and Matrix Factorization Model

Hu¹,

Li²,

Zhang³

2018

dtcse

View full text Add to dashboard Cite

show abstract

“…In [27], the least trimmed squares estimator-based MF (LTSMF) is proposed, which shows better robustness and accuracy compared with MMF. Since LTS-estimator trims ratings with the largest residuals, some genuine users' ratings that have the same or even larger residuals may be trimmed as well, which causes a loss in recommendation accuracy to some extent.…”

Section: Related Workmentioning

confidence: 99%

Robust collaborative recommendation algorithm based on kernel function and Welsch reweighted M‐estimator

Zhang

Sun

2015

IET Information Security

View full text Add to dashboard Cite

The existing collaborative recommendation algorithms based on matrix factorisation (MF) have poor robustness against shilling attacks. To address this problem, in this study the authors propose a robust collaborative recommendation algorithm based on kernel function and Welsch reweighted M-estimator. They first propose a median-based method to calculate user and item biases, which can reduce the influence of shilling attacks on user and item biases because median is insensitive to outliers. Then, they present a method of similarity computation based on kernel function, which can obtain the information of similar users by non-linear inner product operation. Finally, they combine the user and item biases based on median and the similarity based on kernel function with MF model, and introduce the Welsch reweighted M-estimator to realise the robust estimation of user feature matrix and item feature matrix. The experimental results on the MovieLens dataset show that the proposed algorithm outperforms the existing algorithms in terms of both recommendation accuracy and robustness, and the improvement of its robustness is not at the expense of recommendation accuracy.

show abstract

“…But this method only works on moderate attacks. The least trimmed squares estimator based matrix factorization (LTSMF) (Cheng and Hurley 2010) shows better robustness and accuracy compared with MMF. LTSestimator trims part of the largest residuals, which may cause the loss of recommendation accuracy.…”

Section: Introductionmentioning

confidence: 99%

Robust recommendation method based on suspicious users measurement and multidimensional trust

Zhang

2015

J Intell Inf Syst

View full text Add to dashboard Cite

The existing collaborative recommendation algorithms have poor robustness against shilling attacks. To address this problem, in this paper we propose a robust recommendation method based on suspicious users measurement and multidimensional trust. Firstly, we establish the relevance vector machine classifier according to the user profile features to identify and measure the suspicious users in the user rating database. Secondly, we mine the implicit trust relation among users based on the user-item rating data, and construct a reliable multidimensional trust model by integrating the user suspicion information. Finally, we combine the reliable multidimensional trust model, the neighbor model and matrix factorization model to devise a robust recommendation algorithm. The experimental results on the MovieLens dataset show that the proposed method outperforms the existing methods in terms of both recommendation accuracy and robustness.

show abstract

Robust Collaborative Recommendation by Least Trimmed Squares Matrix Factorization

Cited by 25 publications

References 11 publications

Robust Recommendation Method Based on Shilling Attack Detection and Matrix Factorization Model

Robust Recommendation Method Based on Shilling Attack Detection and Matrix Factorization Model

Robust collaborative recommendation algorithm based on kernel function and Welsch reweighted M‐estimator

Robust recommendation method based on suspicious users measurement and multidimensional trust

Contact Info

Product

Resources

About