汉语语料的自动分类

语料库语言学的发展要求语料库的规模越来越大。随着电子出版业的迅速发展, 获取大量机读文本建立大规模语料库已成为可能。但是收集来的粗语料是杂乱无章的, 在作加工整理前必须分类。若用手工分类则工作量很大。本文介绍了一种语料自动分类办法。它采用文中提出的语料相关系数的概念, 并利用不同类语料相关系数不同的特点进行分类, 取得了93%的大类分类正确率。

Automatic Classification of Chinese Texts

he corpus lingUistics requires a very large corpus. Huge amount of computere-reada-ble texts for building very large corpus become available as the rapid development of electronic publishing in recent years.However,the raw texts, which are miscellaneous,must beclassified on domain before further processing.To classify texts is a tedious job for human.Therefore,we present in this paper a new approach of classifying the raw texts automaticaly The related coefficient of texts, which is defined in this paper,could be implied to classify texts.The accuracy rate of classifying texts to 21 classes is 93%.