摘要： 在借鉴汉语基于语料的词抽取技术研究成果的基础上，给出藏语文本预处理方法，并提出一种基于语料库的藏语高频词抽取算法，其中包括藏语文本预处理用噪音字表、紧缩词及其预处理方法和基于语料库的藏语高频词抽取算法。实验结果表明，该算法的准确率达86.22%，召回率达89.79%，F值达87.94%。

关键词: 信息处理, 高频词, 藏语抽词, 自动分词, 语料库, 预处理

Abstract: Based on the research foundlings of Chinese corpus’ extraction, this paper presents the Tibetan preprocessing method and the high-frequency words extraction algorithm, which consists of the tables of noise words, tighten-word, preprocessing method and high-frequency words extraction algorithm. Experimental results show that this algorithm achieves a precision of 86.22%, a recall of 89.79%, and an F-measure of 87.94%.

Key words: information processing, high-frequency words, Tibetan words extraction, automatic words segmentation, corpus, preprocessing

中图分类号: 

• TP391