【作者】 李军辉； 朱巧明； 钱培德；

【Author】 Li Junhui, Zhu Qiaoming, Qian Peide School of Computer Science and Technology, Suzhou University, Suzhou 215006

【机构】 苏州大学计算机科学与技术学院；

【摘要】 本文介绍了在构建动态邮件语料库时所必须解决的问题-邮件语料的添加。由于待添加的邮件预先都已标好类别,根据其类别是否属于原有的类别体系,分两种情况进行处理:1)属于原有类别的新邮件, 将这类邮件细分为Ⅰ、Ⅱ和Ⅲ类,最后只将属于Ⅲ类的新邮件添加到原语料库中;2)不属于原有类别的新邮件,将此类邮件进行去重后全部添加到语料库中。模拟的邮件语料添加实验表明了待添加语料裁剪的必要性及添加算法的有效性。更多还原

【Abstract】 The paper issues an inevitable problem namely Email-adding when constructing a dynamic Email corpus. The newly additive Emails are all class-defined, therefore, according to their labels belonging to the existed classes or not. they are processed in two cases: 1) Provided the new Emails belong to the existed classes, they are classified further as Ⅰ, Ⅱ or Ⅲ. and those belonging to Ⅲ will be added into the corpus. 2) Otherwise, all the Emails are added into the corpus after deleting the repetition. The simulated experiments show the necessity of pruning the new Emails and the validity of Email-adding algorithm.更多还原