Font Size: a A A

Based On The Statistics Of Chinese Word Segmentation In The Mechanical Product Design Of Application

Posted on:2013-08-31Degree:MasterType:Thesis
Country:ChinaCandidate:L JiangFull Text:PDF
GTID:2232330395956212Subject:Mechanical Manufacturing and Automation
Abstract/Summary:PDF Full Text Request
As the fundamental of text mining, machine translation and information retrieval,Chinese word segmentation is a key component in natural language understanding. Dueto the complexity of the Chinese word segmentation itself, accurate and effectivemethod for Chinese word segmentation becomes one of the main studies on naturallanguage understanding. In this thesis, a segmentation system model is establishedbased on statistics and semantic analysis, and is used in product design. This modeltransforms the requirement described in natural language into conceptual designrequirement the computer can recognize.In this paper, the existed methods to segment Chinese words and the main problemsinvolved are analyzed, and the segmentation model based on corpus and its keytechniques are researched. Combined with the present methods, a segmentation modelto combine mechanical segmentation, corpus statistical segmentation with semanticanalysis is proposed, which is then embedded into the whole field Chinese wordunderstanding system. At the stage of mechanical segmentation, the improvedmaximum matching method is used to try to obtain all the segmentation results whichare shown in directed graphs, and the ambiguity character strings are all got. At thestage of corpus statistics, the co-occurrence degree and the collocations of a variety ofcombinations of ambiguity character strings are firstly gathered by corpus statisticsmethod and the collocations are fed back to statistical dictionary, realizing thatsegmentation system and the whole natural language understanding system areself-perfect and in benign interaction. Meanwhile, the credibility of compound words ofthe modes “2+2”,“2+3”and “3+2”is gathered. Then, using gerund structure in semanticanalysis, the segmentation results are further disambiguated combined the conceptualdependency knowledge expression method, improving the accuracy of segmentationsystem and reducing its complexity. Finally, the segmentation model is applied to thefield of user needs analysis of product design with the actual situation.
Keywords/Search Tags:Natural Language, Understanding (NLU), Segmentation, StatisticsProduct Design
PDF Full Text Request
Related items