电子学报

• 科研通信 • 上一篇    下一篇

基于特征选择的推荐系统托攻击检测算法

伍之昂1, 庄毅2, 王有权3, 曹杰1,3   

  1. 1. 南京财经大学江苏省电子商务重点实验室, 江苏南京 210003;
    2. 浙江工商大学计算机与信息工程学院, 浙江杭州 310018;
    3. 南京理工大学计算机科学与技术学院, 江苏南京 210094
  • 收稿日期:2011-09-25 修回日期:2012-02-16 出版日期:2012-08-25
    • 作者简介:
    • 伍之昂 男,1982年生于江苏宜兴,博士,现为南京财经大学江苏省电子商务重点实验室副教授,主要研究领域为推荐系统,云计算和数据挖掘. E-mail:zawu@seu.edu.cn 庄 毅 男,1978年生于浙江杭州,博士,浙江工商大学副教授,获2008中国计算机学会优秀博士论文奖,研究方向为不确定数据管理、多媒体数据库等. E-mail:zhuang@mail.zjgsu.edu.cn
    • 基金资助:
    • 国家自然科学基金 (No.61103229,No.71072172,No.61003074); 浙江省自然科学基金 (No.Z110822,No.Y1110644,No.Y1110969,No.Y1090165); 江苏省科技支撑计划工业部分 (No.BE2011198); 江苏省高等学校优秀科技创新团队 (No.2011013); 东南大学江苏省网络与信息安全重点实验室开放课题 (No.BM2003201); 江苏省高校科研成果产业化推进项目 (No.JHB2011-21)

Shilling Attack Detection Based on Feature Selection for Recommendation Systems

WU Zhi-ang1, ZHUANG Yi2, WANG You-quan3, CAO Jie1,3   

  1. 1. Jiangsu Provincial Key Laboratory of E-Business, Nanjing University of Finance and Economics, Nanjing, Jiangsu 210003, China;
    2. College of Computer and Information Engineering, Zhejiang Gongshang University, Hangzhou, Zhejiang 310018, China;
    3. College of Computer Science and Technology, Nanjing University of Science and Technology, Nanjing, Jiangsu 210094, China
  • Received:2011-09-25 Revised:2012-02-16 Online:2012-08-25 Published:2012-08-25
    • Supported by:
    • National Natural Science Foundation of China (No.61103229, No.71072172, No.61003074); National Natural Science Foundation of Zhejiang Province,  China (No.Z110822, No.Y1110644, No.Y1110969, No.Y1090165); Industry Project of Science and Technology Support Project of Jiangsu Province (No.BE2011198); Excellent Science and Technology Innovation Team of Universities in Jiangsu Province (No.2011013); Open Project of State Key Laboratory of Network and Information Scurity of Southeast University in Jiangsu Province (No.BM2003201); Industrialization Promotion Program of Scientific Research Achievemnet of colleges and universities in Jiangsu Province (No.JHB2011-21)

摘要: 基于协同过滤的电子商务推荐系统极易受到托攻击,托攻击者注入伪造的用户模型增加或减少目标对象的推荐频率,如何检测托攻击是目前推荐系统领域的热点研究课题.分析五种类型托攻击对不同协同过滤算法产生的危害性,提出一种特征选择算法,为不同类型托攻击选取有效的检测指标.基于选择出的指标,提出两种基于监督学习的托攻击检测算法,第一种算法基于朴素贝叶斯分类;第二种算法基于k近邻分类.最后,通过实验验证了特征选择算法的有效性,及两种算法的灵敏性和特效性.

关键词: 推荐系统, 托攻击检测, 特征选择, 朴素贝叶斯分类, k近邻分类

Abstract: Most of the e-business recommender systems are based upon collaborative filtering (CF) algorithms.Since such systems have been shown to be vulnerable to shilling attacks in which malicious user profiles are inserted into the system in order to push or nuke the predictions of some targeted items,shilling attack detection has recently become a hot research topic in recommender systems.Firstly,the effectiveness of five types of attacks against different CF algorithms is analyzed.Secondly,a feature selection algorithm is presented.Two kinds of shilling attack detection algorithms based on supervised learning are then proposed:the first one is based on naÏve Bayesian classifier,and the second one is based on k nearest neighbor (kNN) classifier.At last,experimental results show the effectiveness of the feature selection algorithm and the sensitivity and specificity of these two kinds of detection algorithms.

Key words: recommender system, shilling attack detection, feature selection, naÏ, ve Bayesian classifier, kNN classifier

中图分类号: