Semantic Extraction On Chinese Language Based On Phrase Structure

In the realm of NLP, the syntactic analysis of any language is a difficult and hot problem. We have discovered some strategies to deal with this problem in western languages, such as the generally applicable method: to establish respective syntactical tree for each sentence. However, Chinese language is set up on a loosely constructed grammatical architecture, and, its semantic information is more independent of the grammatical structure. As a result, the syntactic tree of Chinese language offers less accurate and less useful information than western languages.Concerning the characteristic of Chinese language and need of its applications, we propose in this paper an approach and a system which implements it based on the method of extracting phrases from text. It is also shown in this paper that this approach is supported by robust theories. The phrase structure of Chinese language is an extensible organization, through which we can more easily obtain the structure of a sentence. We also can obtain sufficient semantic information from the text, which will satisfy the large-scalar processing of text, after we combine the information obtained from the phrases of a piece of text.Compared with the classic method, which namely is"establishing syntactical tree", the extraction of phrase is configurable with arguments so as to adjust it to do no more work than getting grammatical and semantic information that we need. By means of it, we can decrease the quantity of work that we must do.This method is applicable in broad area. It can be used as former stage for concrete syntactic analysis while we use afterwards this grammatical information directly. The semantic information which is generated by the mean time can also be used. In many circumstance, this semantic information generated in the level of phrase structure is sufficient when the systems are used for automated answering system, sophisticated searching engine, etc.