基于用户行为的个性化邮件分类算法  

Personalized mail classification algorithm based on users' behaviour

作　　者：陈宇[1] 

机构地区：[1]福建工程学院信息科学与工程学院,福建福州350118

出　　处：《福建工程学院学报》2014年第1期83-87,共5页Journal of Fujian University of Technology

摘　　要：针对垃圾邮件过滤过程中分类模型难以个性化、难以适应用户兴趣动态变化的问题,提出了一种基于用户行为的个性化邮件分类算法。算法通过分析朴素贝叶斯分类算法的原理,改造朴素贝叶斯的算法使其具有动态调整能力。实验结果表明此算法具有较高的分类精度。As it is difficult to build a personalized classification model to filter spam to cater to the users＇ changing interests, a novel mail classification algorithm based on users＇ behaviours was pro- posed. With analyses of naive Bayesian algorithm, the computing procedure was rebuilt to possess dynamic adjustment abilities of the classification. Combined with the users＇ operation of the emails in their mailbox, the model for email classification was automatically adjusted to filter the spasm more effectively. The experimental results show that the new algorithm has better precision than the naive Bayesian one.

关 键 词：分类算法 垃圾邮件 朴素贝叶斯 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]