A shilling attack detector based on convolutional neural network for collaborative recommender system in social aware network

Tong, Chao; Yin, Xiang; Li, Jun; Zhu, Tongyu; Lv, Renli; Sun, Liang; Rodrigues, Joel J. P. C.

doi:10.1093/comjnl/bxy008

Cited by 53 publications

(27 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A number of detection techniques exist that reveal unusual or malicious patterns of profile creation and rating data, indicating an attack (Burke et al, 2006). Classical classification techniques (Hurley et al, 2009) such as Bayesian classifier (Wu et al, 2012), peer comparison (OMahony et al, 2005Chirita et al, 2005), but also neural networks are used for detecting deep level features that indicate shilling attacks (Tong et al, 2018). However, the success of classifying the suspicion is dependent on what is defined as malicious in the attack model used to build the classification.…”

Section: Evaluation and Optimizationmentioning

confidence: 99%

The recommender canvas: A model for developing and documenting recommender system design

Capelleveen

Amrit

Yazan

et al. 2019

Expert Systems with Applications

View full text Add to dashboard Cite

The task of designing a recommender system is a complex process. Because of the many technological advancements that may be included in a recommender system, engineers are faced with a fast growing number of design related decisions to be taken. Unfortunately, there is no general approach yet for decision makers that can act as a framework guiding the design of a recommender system. The rich collection of literature on recommender systems, though, offers a great source to identify the key areas where these decisions need to be taken. In this paper, we survey existing literature with the aim of building a recommender system model inspired by Osterwalder's canvas theory. The result of our semi-structured synthesis is a novel design approach in the form of a canvas for designing recommender systems. This work provides a better understanding and can serve as a guide for decision making in recommender system design.

show abstract

Section: Evaluation and Optimizationmentioning

confidence: 99%

The recommender canvas: A model for developing and documenting recommender system design

Capelleveen

Amrit

Yazan

et al. 2019

Expert Systems with Applications

View full text Add to dashboard Cite

show abstract

“…For user-item popularity matrix, we first define the novelty of item and then convert the rating records to user-item popularity matrix by replacing rating value with novelty of item. Moreover, since the sequence of items affects the detection results in deep learning methods [22], the items are sorted by novelty of item in order to cluster the similar items and facilitate the learning of SDAEs.…”

Section: Data Preprocessingmentioning

confidence: 99%

“…(6) CNN-SAD: A method based on deep-level features, which are extracted from users rating profiles by convolutional neural network [22]. According to the cross validation, the vector is reshaped into rectangles with the short side length 20.…”

Section: Experiments On the Movielens 1m Datasetmentioning

confidence: 99%

Detecting Shilling Attacks with Automatic Features from Multiple Views

Hao

Zhang

Wang

et al. 2019

Security and Communication Networks

View full text Add to dashboard Cite

Due to the openness of the recommender systems, the attackers are likely to inject a large number of fake profiles to bias the prediction of such systems. The traditional detection methods mainly rely on the artificial features, which are often extracted from one kind of user-generated information. In these methods, fine-grained interactions between users and items cannot be captured comprehensively, leading to the degradation of detection accuracy under various types of attacks. In this paper, we propose an ensemble detection method based on the automatic features extracted from multiple views. Firstly, to collaboratively discover the shilling profiles, the users’ behaviors are analyzed from multiple views including ratings, item popularity, and user-user graph. Secondly, based on the data preprocessed from multiple views, the stacked denoising autoencoders are used to automatically extract user features with different corruption rates. Moreover, the features extracted from multiple views are effectively combined based on principal component analysis. Finally, according to the features extracted with different corruption rates, the weak classifiers are generated and then integrated to detect attacks. The experimental results on the MovieLens, Netflix, and Amazon datasets indicate that the proposed method can effectively detect various attacks.

show abstract

“…Supervised learning methods allow to detect known attacks, but they are not suitable for combined attacks [7]. To increase the efficiency of detecting attacks of various types, a convolutional neural network with deep learning is proposed in [8]. However, learning such a network requires significant computational costs, which makes it difficult to use it when dynamically changing user interests.…”

Section: Introductionmentioning

confidence: 99%

Method for Detecting Shilling Attacks in E-Commerce Systems Using Weighted Temporal Rules

Chala

Novikova

Chernyshova

2019

EUREKA: Physics and Engineering

View full text Add to dashboard Cite

The problem of shilling attacks detecting in e-commerce systems is considered. The purpose of such attacks is to artificially change the rating of individual goods or services by users in order to increase their sales. A method for detecting shilling attacks based on a comparison of weighted temporal rules for the processes of selecting objects with explicit and implicit feedback from users is proposed. Implicit dependencies are specified through the purchase of goods and services. Explicit feedback is formed through the ratings of these products. The temporal rules are used to describe hidden relationships between the choices of user groups at two consecutive time intervals. The method includes the construction of temporal rules for explicit and implicit feedback, their comparison, as well as the formation of an ordered subset of temporal rules that capture potential shilling attacks. The method imposes restrictions on the input data on sales and ratings, which must be ordered by time or have timestamps. This method can be used in combination with other approaches to detecting shilling attacks. Integration of approaches allows to refine and supplement the existing attack patterns, taking into account the latest changes in user priorities.

show abstract

A shilling attack detector based on convolutional neural network for collaborative recommender system in social aware network

Cited by 53 publications

References 12 publications

The recommender canvas: A model for developing and documenting recommender system design

The recommender canvas: A model for developing and documenting recommender system design

Detecting Shilling Attacks with Automatic Features from Multiple Views

Method for Detecting Shilling Attacks in E-Commerce Systems Using Weighted Temporal Rules

Contact Info

Product

Resources

About