應(yīng)用條件隨機(jī)場(chǎng)進(jìn)行漢語(yǔ)詞法分析、語(yǔ)塊分析的研究.pdf

1、學(xué)校代號(hào)；10731學(xué)號(hào)：P0308120324密級(jí)：公開(kāi)蘭州理工大學(xué)碩士學(xué)位論文應(yīng)用條件隨機(jī)場(chǎng)進(jìn)行漢語(yǔ)詞法分析、語(yǔ)塊分析的研究Y81D291掌僮寶靖厶絲名：星恒昱婭絲名盈驅(qū)整；王繼萱熬攫埴羞皇焦；讓簋也量通信堂睦童些芻整；!鹽簋扭廛旦迨塞握窯旦朔i2Q籩生目一迨塞簽瓣旦朔；2QQ生目21日答趲蠶員金圭顯!堂瑾武塑握應(yīng)用條件隨機(jī)場(chǎng)進(jìn)行漢語(yǔ)詞法分析、語(yǔ)塊分析研究AbstractThisdissenationintroducestheres

2、earchoflexicalanalysisandsymaxparsingisimportant，cmcialand向ndamentalintheresearchofnaturallanguageunderstandingAccordingpresentlythetendencyofmethodsthatintegratestatisticsbasedandmle_basemethods，thispaperintmducestherul

3、esofMaximumEntropya11dthesigni矗canceofitonnatufal1anguageunderstandjngresearchFurthermore，thisdissertationdiscussesthedefinitionandparameterestimateofConditionRandomFieldsCRFsarepmbabiljsticmodelsforsegmemingandlabelings

4、equencedataandheaVilymotivatedbytheprincipleofmaximumentropyThenthisdissertationpresentsaunifiedapproachforChinese1exicalanalysisusingConditjonalRandomFieldsPreciousapplicationsappIyingconditionalrandomfieIdstoChinesewor

5、dssegmentationconve『tsegmemationtochamcterbasedBegin仃nsidetaggingThisdissertationpresemsusingthewordslatticeasthefundamentalsequencetobetaggedtoachieveChineselexicalanalysisThenthe1exiconcanbeusedefficiently：andlanguagek

6、nowledgecanbeintegratedeasilyinfeaturetemplateselectingThisdisseftationalsodiscussesapplyingConditionalRand。mFieldstoChineseChunkParsingandour如tureworksKeywords：Chjneselexicalanalysis；ChineseChunkParsing；CondhionalRandom