Shilling attacks against recommender systems: a comprehensive survey

Güneş, İhsan; Kaleli, Cihan; Bilge, Alper; Polat, Hüseyin

doi:10.1007/s10462-012-9364-9

Cited by 256 publications

(197 citation statements)

References 62 publications

Supporting

Mentioning

183

Contrasting

Unclassified

Order By: Relevance

“…The user profile refers to a collection of user ratings to all items [6]. In order to behave like normal users, the attackers use attack models to create attack profiles based on their knowledge about the recommender systems, such as rating database and item popularity [18].…”

Section: Attack Modelsmentioning

confidence: 99%

“…Existing shilling attack detection methods mostly derived features from rating values in the user profiles [6]. As attackers gain limited knowledge about the systems, their ratings given to items are different from those of normal users [6].…”

Section: Features Derived From Rating Patternsmentioning

confidence: 99%

“…As attackers gain limited knowledge about the systems, their ratings given to items are different from those of normal users [6]. Among these features, DegSim and RDMA are the most popular ones and are calculated according to the formulas (1) and (2) [6], [21], [22].…”

Section: Features Derived From Rating Patternsmentioning

confidence: 99%

“…2.1, there are three kinds of attacks when only selecting patterns are considered [6], [7]: 1) Attackers who randomly choose items to rate from the whole item set, such as random attacks and average attacks; 2) Attackers who randomly choose items to rate from the top-x% popular item set, such as average over popular attacks (Aop attacks); 3) Attackers who randomly choose items to rate both from the whole item set and from certain kind of item set, such as bandwagon attacks. Other attackers can be included in these three categories.…”

Section: Popularity Characteristics Of User Profilesmentioning

confidence: 99%

“…The purpose of shilling attacks is to manipulate the recommendation results, therefore, attackers need some strategies to select items to rate and give rating values to items, which means selecting patterns and rating patterns respectively. Randomly inserted profiles without attack models will cause no significant impact on the recommender systems [6]. But attackers' knowledge about the recommender systems is limited [13].…”

Section: Effectiveness Analysismentioning

confidence: 99%

See 4 more Smart Citations

Shilling Attack Detection in Recommender Systems via Selecting Patterns Analysis

Gao

Hua

et al. 2016

IEICE Trans. Inf. & Syst.

View full text Add to dashboard Cite

SUMMARYCollaborative filtering (CF) has been widely used in recommender systems to generate personalized recommendations. However, recommender systems using CF are vulnerable to shilling attacks, in which attackers inject fake profiles to manipulate recommendation results. Thus, shilling attacks pose a threat to the credibility of recommender systems. Previous studies mainly derive features from characteristics of item ratings in user profiles to detect attackers, but the methods suffer from low accuracy when attackers adopt new rating patterns. To overcome this drawback, we derive features from properties of item popularity in user profiles, which are determined by users' different selecting patterns. This feature extraction method is based on the prior knowledge that attackers select items to rate with man-made rules while normal users do this according to their inner preferences. Then, machine learning classification approaches are exploited to make use of these features to detect and remove attackers. Experiment results on the MovieLens dataset and Amazon review dataset show that our proposed method improves detection performance. In addition, the results justify the practical value of features derived from selecting patterns.

show abstract

Section: Attack Modelsmentioning

confidence: 99%

Section: Features Derived From Rating Patternsmentioning

confidence: 99%

Section: Features Derived From Rating Patternsmentioning

confidence: 99%

Section: Popularity Characteristics Of User Profilesmentioning

confidence: 99%

Section: Effectiveness Analysismentioning

confidence: 99%

See 3 more Smart Citations

Shilling Attack Detection in Recommender Systems via Selecting Patterns Analysis

Gao

Hua

et al. 2016

IEICE Trans. Inf. & Syst.

View full text Add to dashboard Cite

show abstract

Recommendations and privacy in the arXiv system: A simulation experiment using historical data

Menkov

Ginsparg

Kantor

2019

Asso for Info Science & Tech

View full text Add to dashboard Cite

Recommender systems may accelerate knowledge discovery in many fields. However, their users may be competitors guarding their ideas before publication or for other reasons. We describe a simulation experiment to assess user privacy against targeted attacks, modeling recommendations based on co-access data. The analysis uses an unusually long (14 years) set of anonymized historical data on user-item accesses. We introduce the notions of "visibility" and "discoverability." We find, based on historical data, that the majority of the actions of arXiv users would be potentially "visible" under targeted attack. However, "discoverability," which incorporates the difficulty of actually seeing a "visible" effect, is very much lower for nearly all users. We consider the effect of changes to the settings of the recommender algorithm on the visibility and discoverability of user actions and propose mitigation strategies that reduce both measures of risk.

show abstract

Towards the significance of taxi recommender systems in smart cities

Katarya

2022

Concurrency and Computation

View full text Add to dashboard Cite

Since their launch in the early 1990's, recommender systems (RSs) have played an essential role in information filtering and providing personalized information to users by utilizing the web, mobile, context, personalized, and practical applications. Many research articles have been published in various application domains such as movies, books, documents, news, images, music, shopping, TV programs, tourism, group, social, taxi, and others. However, from these domains, taxi recommendation is a new and thrust area of research, as publications in this domain are few and recently published.Taxi recommendations are associated with drivers-passengers with their: behavior, hotspots, trajectory, point of interest, route planning or prediction, taxi finding, anomaly or fraud detection, urban computing, mobility patterns and so forth. Due to this diversity and various agendas in taxi recommendations, more exploration is required, as it is a less mature and growing field than other domains. So we have done a literature review in the taxi domain by classification, incorporating reputed articles published from 2001 to August 2022. We include the reputed papers which highlight, analyze and perform studies only on the taxi domain. This article categorizes and discusses the research and development in taxi recommendations as nature, approaches, algorithms, methods, datasets, models, devices, information, and evaluation techniques. This survey article will actively help the researchers and professionals understand the current situations in taxi RS and resolve future trends, opportunities, and research.

show abstract

Shilling attacks against recommender systems: a comprehensive survey

Cited by 256 publications

References 62 publications

Shilling Attack Detection in Recommender Systems via Selecting Patterns Analysis

Shilling Attack Detection in Recommender Systems via Selecting Patterns Analysis

Recommendations and privacy in the arXiv system: A simulation experiment using historical data

Towards the significance of taxi recommender systems in smart cities

Contact Info

Product

Resources

About