邮件语料库的语料添加算法研究与实现.pdf

邮件语料库的语料添加算法研究与实现+ 李军辉，朱巧明，钱培德 苏州大学计算机科学与技术学院苏州215006 Emaih jhli@zhhz．org 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一邮件语料的添加。由于待添加的邮件预先 都已标好类别，根据其类剐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lII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科库中．模拟的邮件语抖添加实验表明了待添加语料链剪的必 璺性及添加算法的有效性。 关键字：邮件语料库，语料添加，外围语料 The of in ImplementationEmail-AddingAlgorithm anEmail ConstructingCorpus Li Peide Junhui，Zhu Qiaoming，Qian School Scienceand 215006 ofComputerTechnology，SuzhouUniversity，Suzhou issuesallinevitable when n Email Abstract：The dynamic paper problemnamelyEmail-addingconstructing are totheirlabels tothe additiveEmailsall corpus．Thenewly class-defined，therefore，accordingbelonging or are intwo thenewEmails totheexisted existedclassesnoL canes：¨Provided classes， they processed belong be the aleclassifiedfiartheras or those to111will addedinto they I。II 111，and belonging oorpus2) theEmailsaleaddedintothe after the simulated show Otherwise,all corpusdeletingrepetition．Theexperiments the thenewEmailsandthe necessityofpruning validityofEmail·addingalgorithm． words：Email in Key corp