電子商務推薦攻擊研究
發(fā)布時間:2018-07-25 10:23
【摘要】:個性化推薦是實現(xiàn)客戶關系管理的重要手段和技術。協(xié)同過濾作為最核心、最典型的個性化推薦技術,被廣泛應用于電子商務,但其推薦結果對用戶偏好信息敏感,使得推薦系統(tǒng)易受到人為攻擊,電子商務推薦安全成為個性化推薦能否成功應用的關鍵。作者先簡要介紹了電子商務個性化推薦的基本概念,然后系統(tǒng)闡述了推薦攻擊的概念、特征、攻擊成本及攻擊效率,并詳細比較了各種攻擊模型,以及各種攻擊模型對不同推薦模型的穩(wěn)定性和健壯性的影響,分析比較了各種攻擊檢測模型。最后總結評述了電子商務推薦安全的研究現(xiàn)狀,并提出了未來研究的挑戰(zhàn)。
[Abstract]:Personalized recommendation is an important means and technology for the realization of customer relationship management. As the core and the most typical personalized recommendation technology, collaborative filtering is widely used in e-commerce. However, the recommended results are sensitive to user preference information, making the recommendation system vulnerable to human attack. The recommendation security of e-commerce becomes a personalized recommendation. The key to successful application is to introduce the basic concept of personalized recommendation in e-commerce. Then the concept, characteristics, attack cost and attack efficiency of the recommended attack are systematically expounded, and various attack models are compared in detail, and the effects of various attack models on the stability and robustness of different recommendation models are analyzed and compared. Various attack detection models are presented. Finally, the research status of recommendation security in e-commerce is summarized and the challenges of future research are proposed.
【作者單位】: 中國人民大學信息學院 中國人民大學信息學院 中國人民大學信息學院
【基金】:信息管理與信息經濟學教育部重點實驗室開放基金資助(F0607-31)
【分類號】:TP393.08
[Abstract]:Personalized recommendation is an important means and technology for the realization of customer relationship management. As the core and the most typical personalized recommendation technology, collaborative filtering is widely used in e-commerce. However, the recommended results are sensitive to user preference information, making the recommendation system vulnerable to human attack. The recommendation security of e-commerce becomes a personalized recommendation. The key to successful application is to introduce the basic concept of personalized recommendation in e-commerce. Then the concept, characteristics, attack cost and attack efficiency of the recommended attack are systematically expounded, and various attack models are compared in detail, and the effects of various attack models on the stability and robustness of different recommendation models are analyzed and compared. Various attack detection models are presented. Finally, the research status of recommendation security in e-commerce is summarized and the challenges of future research are proposed.
【作者單位】: 中國人民大學信息學院 中國人民大學信息學院 中國人民大學信息學院
【基金】:信息管理與信息經濟學教育部重點實驗室開放基金資助(F0607-31)
【分類號】:TP393.08
【參考文獻】
相關期刊論文 前3條
1 趙亮,胡乃靜,張守志;個性化推薦算法設計[J];計算機研究與發(fā)展;2002年08期
2 余力,劉魯;電子商務個性化推薦研究[J];計算機集成制造系統(tǒng);2004年10期
3 余力,劉魯,羅掌華;我國電子商務推薦策略的比較分析[J];系統(tǒng)工程理論與實踐;2004年08期
【共引文獻】
相關期刊論文 前10條
1 覃遵躍;在左邊是單屬性的函數(shù)依賴集中尋找關系模式候選碼的算法[J];安慶師范學院學報(自然科學版);2003年02期
2 張友志;程玉勝;王一賓;;基于Web日志挖掘的Markov預測模型及算法研究[J];安慶師范學院學報(自然科學版);2010年01期
3 牟乃夏;劉文寶;張靈先;孫翠羽;;空間信息服務的個性化問題[J];測繪科學;2011年03期
4 戰(zhàn)坤;曾凡;康運生;戴黎陽;;個性化——醫(yī)院網站信息服務的趨勢[J];重慶醫(yī)學;2009年21期
5 曹毅;羅新星;;電子商務推薦系統(tǒng)關鍵技術研究[J];湘南學院學報;2008年05期
6 廖華;;基于客戶生命周期與交易偏好的商品推薦方法[J];大家;2010年08期
7 高e,
本文編號:2143524
本文鏈接:http://sikaile.net/guanlilunwen/kehuguanxiguanli/2143524.html
最近更新
教材專著