基于粗糙集和朴素贝叶斯的垃圾邮件过滤系统  

Spam Filtering System Based on Rough Set and Naive Bayes

作　　者：陈炼[1] 王芸[2] 饶泓[1] 邓少波[2] 

机构地区：[1]南昌大学计算中心,江西南昌330031 [2]南昌工程学院计算机科学与技术系,江西南昌330099

出　　处：《南昌大学学报（工科版）》2009年第1期45-48,共4页Journal of Nanchang University(Engineering & Technology)

基　　金：江西省教育厅科技计划资助项目(赣教技字[2007]23号;赣教技字[2007]344号)

摘　　要：提出了基于粗糙集理论和贝叶斯分类算法的垃圾邮件过滤方法。利用粗糙集约简算法对邮件样本集进行特征约简,删除对邮件过滤结果影响不大的冗余特征,从而降低了输入样本集的维数,解决了贝叶斯分类器训练时间长,样本集占用的存储空间过大的问题。实验证明,该方法可以提高邮件过滤的准确性和训练的速度。This paper proposed a spam filtering method based on Rough set theory and Bayesian classifier algo-rithm. Then the amount of features are reduced by deleting redundant features with little significance on filtering effect based on rough set theory, resulting in a input sample with reduced number of dimension. Using this method, it can overcome the shortages of Bayies classifier-time-consuming of training and massive dataset storage. Experiments proved that this mechanism could greatly boost both the system＇ s accuracy and the training speed.

关 键 词：垃圾邮件 粗糙集 朴素贝叶斯分类器 

分 类 号：TP311.5[自动化与计算机技术—计算机软件与理论]