βp: A novel approach to filter out malicious rating profiles from recommender systems

NCU Institutional Repository > 管理學院 > 企業管理學系 > 期刊論文 > Item 987654321/103275

jsp.display-item.identifier=請使用永久網址來引用或連結此文件: https://ir.lib.ncu.edu.tw/handle/987654321/103275

题名:	βp: A novel approach to filter out malicious rating profiles from recommender systems
作者:	許秉瑜;Chung, Chen-Yao;Hsu, Ping-Yu;Huang, Shih-Hsiang
贡献者:	管理學院企業管理學系
关键词:	Algorithms;Applied sciences;Collaborative filtering;Computer science;control theory;systems;Computer systems and distributed systems. User interface;Data processing. List processing. Character string processing;Exact sciences and technology;Filtering;Filtration;Inventory control, production control. Distribution;Memory and file management (including protection and security);Memory organisation. Data processing;Operational research and scientific management;Operational research. Management science;Purchasing;Ratings;Recommender systems;Shilling attacks detection;Software;Statistical process control;Trains;Websites
日期:	2013-04-01
上传时间:	2026-04-23 11:27:04 (UTC+8)
出版者:	Elsevier;Amsterdam: Elsevier B.V
摘要:	摘要： Recommender systems are widely deployed to provide user purchasing suggestion on eCommerce websites. The technology that has been adopted by most recommender systems is collaborative filtering. However, with the open nature of collaborative filtering recommender systems, they suffer significant vulnerabilities from being attacked by malicious raters, who inject profiles consisting of biased ratings. In recent years, several attack detection algorithms have been proposed to handle the issue. Unfortunately, their applications are restricted by various constraints. PCA-based methods while having good performance on paper, still suffer from missing values that plague most user–item matrixes. Classification-based methods require balanced numbers of attacks and normal profiles to train the classifiers. The detector based on SPC (Statistical Process Control) assumes that the rating probability distribution for each item is known in advance. In this research, Beta-Protection (βP) is proposed to alleviate the problem without the abovementioned constraints. βP grounds its theoretical foundation on Beta distribution for easy computation and has stable performance when experimenting with data derived from the public websites of MovieLens. ► Recommender systems may be injected with malicious profiles. ► Prior works eliminated attacking profiles with classification or PCA based methods. ► They either suffer from the lack of negative cases or cannot cope with sparse data. ► To avoid both issues, the work proposed a method based on beta distributions. ► The experiment shows that the proposed method outperforms the others. 出版者： Amsterdam: Elsevier B.V 出版日期： 2013-04-01 出處： Decision Support Systems, 2013-04, Vol.55 (1), p.314-325 資源來源： Elsevier ScienceDirect 版權： 2013 Elsevier B.V. 版權： 2014 INIST-CNRS 識別號： ISSN: 0167-9236 識別號： DOI: 10.1016/j.dss.2013.01.020
显示于类别:	[企業管理學系] 期刊論文

文件中的档案:

档案	描述	大小	格式	浏览次数
index.html		0Kb	HTML	24	检视/开启

在NCUIR中所有的数据项都受到原著作权保护.

社群 sharing

数据加载中.....