中文电子邮件作者的身份判别

Authorship identification of Chinese E-mail

电子邮件已成为因特网上最基本、最重要的应用之一。但利用电子邮件进行诈骗、反动宣传等犯罪现象也日益严重。因此采用研究中文电子邮件作者身份挖掘的方法,以识别邮件作者的真实身份,为计算机取证提供依据。通过分析邮件作者的语言特征、结构特征和格式特征,利用支持向量机算法,自动把邮件文档分类到预定的作者类别中,并对有限数据集的试验取得了满意的结果。

E-mail has become one of the most important application on the lnternet. But, the phenomenon of utilizing e-mail to crime is serious day by day, such as antisocial mail, fraud mail, racketeering mail, terroristic threatening mail and so on. So the paper studies the method mining the Chinese E-mail author＇ s true identity, offers basis for Computer Forensics. We use various e-mail document features to classify authorship of emails such as structural characteristics, linguistic evidence and form characteristics with the Support Vector Machine as the learning algorithm. Experiments on a number of e-mail documents give promising results with some e-mail document features and author categories giving better categorization performance results.