电子邮件分类中的特征选择

Feature Selection in E-mail Classification

电子邮件是互联网的最重要应用之一,尽管给人们日常工作和生活带来很大便利,但也带来了一种令人讨厌的副产品——垃圾邮件。对邮件进行分类已成为当前的一个研究热点,而如何进行邮件特征选择,是邮件分类中一个基本也是很重要的问题。本文在分析比较几种用于邮件分类的典型特征选择方法基础上,提出一种新的结合了 Mi-tra’s算法和顺序前进搜索法优点的邮件特征选择方法。实验结果表明该方法能够改进邮件分类的准确率,验证了本文方法的有效性和可行性。

E-mail is one of the most popular services of the Internet, E-mail has brought us great convenience in our daily work and life. It has brought us an annoying byproduct Spare. How to classify incoming E-mails and filter spam has attracted much attention. One fundamental yet important issue in E-mail classification is how to select the ap propriate features. Based on the analysis and eomparison of several typical feature seleetion methods for E-mail classification, a new method is proposed, which combines both Mitra＇s and Sequential Forward Selection. Experimental result shows that the proposed method can improve the precision of E-mail classification.