基于向量空间模型的古汉语词义自动消歧研究

o explain sense of the words is the important part of the arrangement of Chinese ancient books. Manual interpretation is very time-consuming. Learning from the research of modern word sense disambiguation, an improved unsupervised word sense disambiguation method of the ancient Chinese was proposed based on vector space model of senses. In this article, the knowledge base of the senses of the polysemous words in the ancient Chinese was build, and the contexts and the senses of the polysemous words were mapped to the vector space model in order to complete the task of word sense disambiguation. It was full-text database of ancient Chinese agricultural books that used as statistics corpus, and the word sense tagging experiments were taken for ten typical polysemous words of the ancient Chinese, a total of 29 senses, 1836 contexts. The results showed that the average accuracy of word sense disambiguation was 79.5%.