A Robust Collaborative Recommendation Algorithm Based on Least Median Squares Estimator

Zhang, Fuzhi; Sun, Shuangxia

doi:10.4304/jcp.9.2.308-314

Cited by 7 publications

(7 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…(4) LMedSMF: A robust collaborative recommendation algorithm based on least median squares estimator [36]. As shown in Fig.…”

Section: Experimental Results and Analysismentioning

confidence: 99%

“…In our recent work [36], we proposed a robust collaborative recommendation algorithm based on least median squares estimator (LMedSMF). LMedSMF algorithm introduces LMedS-estimator to filter out the ratings with the largest residuals and uses RLS-estimator to realise the robust estimation of user and item feature matrices, thus improving the robustness and recommendation accuracy to some extent.…”

Section: Related Workmentioning

confidence: 99%

“…Since the measures based on the RBF kernels are robust [37], the outliers cannot cause too much effect on the similarity calculation of KWMMF. The robust parameter estimation methods used in the literatures [27,34,36] suffer in terms of accuracy, which means that the robustness of these methods comes at a cost of accuracy. In KWMMF algorithm, we introduce Welsch reweighted M-estimator to realise the robust parameter estimation, which need not trim outliers in the process of parameter estimation.…”

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

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

Zhang

Sun

2015

IET Information Security

Self Cite

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

“…(4) LMedSMF: A robust collaborative recommendation algorithm based on least median squares estimator [36]. As shown in Fig.…”

Section: Experimental Results and Analysismentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

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

Zhang

Sun

2015

IET Information Security

Self Cite

View full text Add to dashboard Cite

show abstract

“…Recently, an article proposed a Robust RS that resists random attacks and average attacks, which looks like the average user but uses the Filler attack target. Zhang and Sun and Rodríguez‐Hernández and Ilarri, used RDMA and DegSim metrics to detect both general users and other users. In the algorithm of this article, if an item is judged as having received a sybil attack, all users of the project will be deleted .…”

Section: Related Researchmentioning

confidence: 99%

A robust recommended system based on attack detection

Yang

2018

Concurrency and Computation

View full text Add to dashboard Cite

In the increasingly fierce competition in e-commerce sites, the recommendation system has brought great benefits to the site, but some unscrupulous businesses use the recommended system algorithm loopholes, the use of bulk injection of some fake users, and the ratings of these users with the normal user's rating. Therefore, when calculating user similarity, it is easy to enter the user ′ s neighborhood circle, because the false user takes a high score ("push attack") or a low score ("null attack") on the target project. The recommendation scores will be biased, and it is important to detect these false users for the recommendation system and maintain a good e-commerce competitive environment. In this paper, we propose a recommendation algorithm that divides user-generated ratings into normal ratings and non-generic ratings and refine and use state information to minimize the impact of spurious malicious users. Our algorithm first ensures that the recommender system is stable against the three main attack modes (random attack, average attack, power flow attack). Through the analysis of the real data, we verify the performance of the proposed scheme and compare our algorithm with the existing one. KEYWORDSattack model, collaborative filtering, nuclear attack, push attack, recommendation system INTRODUCTIONThe recommendation system is an intelligent software tool for providing advice to users and has been used in many fields such as e-commerce, movie and video websites, music network radio stations, social networks, as well as personalized reading, mailing and, advertising. 1-3In the increasingly fierce competition in e-commerce sites, the recommendation system has brought great benefits to the site, but some unscrupulous businesses use the recommended system algorithm loopholes, the use of bulk injection of some fake users and the ratings of these users with the normal user's rating. Similarly, the user's proximity circle is easily calculated when calculating the user's similarity level. Since the fake user takes a high score ("push attack") or a low score ("nuclear attack") on the target item because the normal user pair The target project's recommended rating will be biased to detect these fake users in terms of the recommendation system is very important to maintain a good competitive environment for e-commerce. If e-commerce sites are often malicious, fake users fake rating; the recommendation system is very difficult to recommend suitable items to the user.Motivatedly, in order to make their products sell well, manufacturers always want recommender systems to frequently recommend their own products while reducing or not recommending competitors' products. Instead of trying to improve the quality of their products, some bad producers do this by using cheating techniques to increase the frequency with which recommender systems recommend their products.The recommendation system is based on similar users or similar products to generate a list of recommendations, but it is also an open system that requires user par...

show abstract

“…However, when a similar user gives unsuccessful advice, there is no way to ignore it. Over and over again, the system will cause a descent in the performance [21]. Therefore, in this paper, trust networks, credibility and trust relationships are adopted to overcome it.…”

Section: Notation and Definitionsmentioning

confidence: 99%

A Trusted Recommendation Method Based on Social Circles

Cheng-jian¹,

Chen²,

Song³

2016

Proceedings of the 2016 International Conference on Computer Science and Electronic Technology

View full text Add to dashboard Cite

Abstract:The crisis of confidence has become increasingly rigorous, especially in recommendation domain. Deception and malicious attacks are found among the traditional collaborative filtering recommendation systems. Taking into account the credibility, the behaviour-based trust relationship, and the social circle-based trust relationship, this paper proposes a trusted recommendation method via social circle to enhance the credibility of recommendation. By using the social circles, which can significantly help to deal with the sparsity and trust problem of the traditional collaborative filtering, the method is aimed to solve two fundamental problems: accuracy and robustness. Empirical evaluation of the method on the dataset of last.fm demonstrates that the method can enhance the robustness without reducing the accuracy of the recommendation.

show abstract

A Robust Collaborative Recommendation Algorithm Based on Least Median Squares Estimator

Cited by 7 publications

References 17 publications

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

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

A robust recommended system based on attack detection

A Trusted Recommendation Method Based on Social Circles

Contact Info

Product

Resources

About