软件导刊2017，Vol.16Issue(2)：42-47,6.DOI:10.11907/rjdk.162568
基于特征分析的推荐系统托攻击检测算法研究
Shilling Attack Detection Algorithm Based on Feature Extraction for Recommendation Systems
摘要
Abstract
Collaborative filtering is one pupularly used recommendation technique in recommender systems.However,reco-mmender systems based on collaborative filtering are highly vulnerable to what have been termed "shilling" or "profile inj-ection" attacks.Attackers inject a number of organized biased ratings in order to manipulate recommendation systems.In or-der to detect the shilling attack users,this paper analyze the statistical features of attackers detailedly and then propose anattack detection algorithm based on the feature analysis.The experimental results show that the proposed attack detection al-gorithm could achieve higher detection rate.Effectively alleviate the problem of the shilling attack to recommendation syste-ms and ensure the reliability of the recommendation systems.关键词
推荐系统/协同过滤/托攻击/特征分析/攻击检测算法Key words
Recommended System/Collaborative Filtering/Shilling Attack/Feature Analysis/Attack Detection Algorithm分类
信息技术与安全科学引用本文复制引用
胡德敏,朱德福..基于特征分析的推荐系统托攻击检测算法研究[J].软件导刊,2017,16(2):42-47,6.基金项目
国家自然科学基金项目(61170277,61472256) （61170277,61472256）
上海市教委科研创新重点项目(12zz137) （12zz137）
上海市一流学科建设项目(S1201YLXK) （S1201YLXK）
