个性化推荐系统安全防护研究  

Research on Security Protection of Personalized Recommendation System

作　　者：李洺吉 王晶[1] Li Mingji;Wang Jing(Beijing University of Posts and Telecommunications, Beijing 100876, China)

机构地区：[1]北京邮电大学网络与交换技术国家重点实验室,北京100876

出　　处：《信息通信技术》2016年第6期43-47,67,共6页Information and communications Technologies

摘　　要：近年来,个性化推荐系统作为电子商务技术的一项重要研究内容,有效地解决了信息过载的问题,为用户提供了精准的个性化推荐服务并为厂商带来了更高的收益。但与此同时,随着数据规模的增大、用户信息数据重要性的提升,推荐系统安全问题开始受到广泛关注。个性化推荐系统如何保证用户数据安全,即使在恶意数据的干扰下仍能保证推荐系统的准确可靠亟待解决。文章就此,分析国内外的研究情况,首先介绍个性化推荐系统及其信任度的概念,明确了用户隐私数据安全对推荐系统的重要性;接着,介绍了常见的用户隐私数据保护策略;最后对推荐系统托攻击与其攻击检测算法进行归纳总结,并提出可供参考的解决方法。In recent years, personalized recommendation system, as one of the important research issues of e-business technology, has effectively solved the “information-overload” problem and provides accurate personalized recommendation service and gains higher benefit at the same time. But with the constant expanding of the data scale and increasing importance of user information, the security problem of recommendation system has become a hotspot. The question that how personalized recommendation system can provide credible recommendation results and protect users’ privacyeven interfeved bymalicious data has become an import urgent problem now. This paper analyzes the research situation both at home and abroad. First,this paper explains the concept of the personalized recommendation system and its credibility, and emphasizes the importance of user privacy information in recommendation system, then introduces some common user privacy protection policies. Finally based on shilling attack and attack detection algorithm, this paper makes an inductive generalization and provides some solutions for reference.

关 键 词：推荐系统 用户隐私 数据安全 个性化 

分 类 号：TP391.3[自动化与计算机技术—计算机应用技术] TP309[自动化与计算机技术—计算机科学与技术]