Analysis Of Natural Language Understanding And Its Application

This article discusses the Natural Language Understanding. It is a high level technical about language and character disposal, and it is an important content and part of artificial intelligence. It researches the theory and method which can help the natural communications between computer and people. This article bases on the Hierarchical Network of Concept. Then, I analyze the concepts and targets about Natural Language Understanding. Lastly, I try to do an experiment about voice identify technical: how to identify dialects. The main thoughts are as follows.1. I found a logical thinking and language module which can help understanding the understanding of the natural language. But according to contemporary science developing level, we cannot easily build this module, and the computer can only approximately simulate people's thinking. Language module should base on the children's language cognition.2. According to traditional linguistics, language is the only tools to communicate and the only carrier of communication. But language cannot successfully fulfill expression and understanding after we analyze the real expression and understanding process. Usually, the tools that we used for expression and understanding are: look, listen, smell, say and feel. If we want to know the real meaning of Language, we must stand at the broad sense. Many linguists say that analyzing and understanding of language is an administration level. When the new information comes up, they come into our cerebra through the five tools. And then they join with the old information after the arrangement. Just like telegraph has a fixed reference, new information can understand only through the combination with the old information.3. If the machine language really simulates natural language, it should have five ultimate qualifications. (l)Machine language can input real understanding with no restrict. (2)Machine language can carry out information background which can match to people's one. (3) Machine language can carry out psychic and consciences structure which is similar to people. (4) Machine language act by the way which is the same with people's. (5) Machine language can output by the way which is the same with people's.4. Abundant dialect resource is the important part of the Chinese culture, so protecting dialect is a formidable task. Founding a dialect data bank, registering several dialects sound, rhyme and expression are linguists' important task. This article try to design a simple dialect identifying module, starting by identify Xiangfan dialect. According to Corpus Linguistics, I coordinate characteristic words in Xiangfan dialect, then input these words into computer and last label them. When we should distinguish a paragraph, we can use this corpus.