Font Size: a A A

The Application Of The Noun Cluster Of NLU In The Domain Of Mechanical Design

Posted on:2010-05-11Degree:MasterType:Thesis
Country:ChinaCandidate:X D FeiFull Text:PDF
GTID:2178360272482550Subject:Mechanical Manufacturing and Automation
Abstract/Summary:PDF Full Text Request
This paper described the application of the Noun Cluster of NLU in the domain ofmechanical product design, did more research into the methods and realization of nouncluster in discourse. Finally, understood and analyzed the user's need expressed in theform of natural language, and transformed analysis result into the design requirement ofconcept or design parameter, offered a good design environment to users and designers.The main work done in this paper is as following:Several theories and methods of NLU knowledge representation was introduced,based on this, got a new method of knowledge representation used "concept"including its property, intension and extension as the kernel, based on frameworkstructure, expressed by hierarchy structure like a tree showing the association ofconcepts. In the end, the method was used to construct a knowledge database and somerules. Semanteme-understanding and cluster-processing moulds of noun wereestablished combined with the typical problems in noun-processing;also the core-word theory was presented. Two important methods in noun cluster were given, one issequencing to mark of dynamic noun and the other is chain-searching based onsemanteme-understanding,expressed their basic thought and accomplished course, andrealized the methods using examples. Also this thought was applied to solve problemsof ambiguity produced by default words. At last, the NLU based on domain and thenoun cluster of discourse was applied to the analysis of user's needs in mechanicaldesign expert system. Meanwhile, the knowledge database and some rules wereestablished, all these work done came to a good result.
Keywords/Search Tags:Natural Language Understanding, Discourse Knowledge Database, Noun Cluster, Product Design
PDF Full Text Request
Related items