基于P2P网络的个性化协同邮件过滤模型  

Personalized Collaborative E-mail Filtering Model Based on P2P Network

作　　者：司广涛[1] 孟静[1] 李光顺[1] 

机构地区：[1]曲阜师范大学计算机科学学院,山东日照276826

出　　处：《计算机技术与发展》2011年第6期128-131,共4页Computer Technology and Development

基　　金：山东省优秀中青年科学家科研奖励基金(BS2009DX024);曲阜师范大学科研基金(XJ0703;XJ200809)

摘　　要：基于内容的协同过滤器利用用户关于误判和漏判信息的反馈,对电子邮件进行过滤,可以提高过滤器跟踪垃圾邮件中概念漂移的能力,但这类方法都对集中式的垃圾邮件过滤做出了隐含假设。为了解决集中式协同邮件过滤的问题,提出了一种基于P2P网络的个性化协同垃圾邮件过滤模型,描述了其系统结构和基于签名的系统设计与实现方法。实现了一个原型系统,收集使用其作为垃圾邮件过滤器的用户的数据进行分析。实验结果分析表明,这种个性化协同垃圾邮件过滤是准确和有效的。E-mail is filtered with user＇s feedback information about false positives and false negatives in content-based collaborative filters.While this improves the ability of the filters to track concept drift in spam over time,such approaches make implicit assumptions in centralized spam filtering.A personalized collaborative spam filtering model based on P2P network was presented to solve the problems in centralized collaborative spam filters and the design and implementation of signature-based system and the architecture were described.A prototype system was implemented to analyze the data collected from the users employing the system as their spam filtering tool.The evaluation of the results indicates that the personalized collaborative spam filtering is accurate and efficient.

关 键 词：P2P网络 个性化 协同 垃圾邮件过滤 

分 类 号：TP393.2[自动化与计算机技术—计算机应用技术]