The Automatic Recognition Research On Chinese Modality's Usage Based On Rules And Statistical

Researching on automatic recognizing usages of modern Chinese modalities is one of the important contents of the NLP-oriented Chinese modalities knowledge base. This paper try to automatic recognize Chinese modality's usages with the methods based rules and statistic. First, this paper built the modality's usage dictionary. The relation between the auxiliary and the modality is very close, if we want to achieve the automatic recognition of the auxiliary and the modality, it should be identify word parts of speech first. So this paper research and identify the word POS of some controversial words.Secondly, this paper recognized the modality's usages with the modality's usage rules base, and studied the automatic improvement of rules based on error-driven by way of extract additional rules for the reason that the rules written by manual are subjectively and can hardly cover all usages in true text.Thirdly, this paper study automatically recognizing Chinese modality's usages using statistical methods. Finally, this paper combined the methods of rule-based and statistical-based. The experiments show that the method combined rule-based and statistical-based is effective in automatically recognizing of modality's usages.In the end, this paper summarizes the research word, the next research forecasted, and prospects about Chinese modality's usages.