基于词二元接续的中文文本自动查错研究

The Research of Chinese Text Automatic Error-Checking Method Based on the Neighboring Relations of Words

运用语料库语言学统计方法对中文文本自动查错的有关问题进行探讨 ,运用词二元接续关系进行查错 ,主要依据词二元同现概率、互信息、t -测试差 .其中 ,t-测试差是首次被应用于查错 .

In this paper, the statistical methods of corpus linguistics are applied to solve the problem of checking. And when checking, the relations between words are considered. When the relations between words are concerned, the bi-gram co-occurrence probability, mutual information and the difference of t-test are considered.