中文信息處理中若干技術的研究與實現(xiàn)

1、遼寧科技大學碩士學位論文中文信息處理中若干技術的研究與實現(xiàn)姓名：羅麗俊申請學位級別：碩士專業(yè)：計算機應用技術指導教師：戰(zhàn)學剛20080322遼寧科技大學碩士論文AbstractWiththerapiddevelopmentoftheInteract，allkindsofresourcesarecontinuouslyincreasingInordertoquicklyandefficientlyfindtheinformation，in

2、formationprocessing(IP)hasbecomeoneofthemostimportantresearchfieldsInthisthesis，wediscusssomekeyissuesinChineseinformationprocessingThemajorcomeⅡtSandcontributionsareasfollows：Implementationofadictionarybasedonsyntacticp

3、arsingmethodsBymappinggrammarrulesforthesignatureword，parsingisconvertedtothegenerationofcharacteristicsofsyntacticdecisiontreebytheusetheterm，SOthattherule—basedmethodandprobabilitymethodareeffectivelycombinedInourclose

4、dtest，theprecisionandrecallare8713％and8940％，respectivelyAnimprovementoftheK—meansclusteringmethodhasbeenproposedbytheuseofsampledistanceThiseffectivelyavoidserrorscausedbythechoiceoftheinitialpointaswellasnoiseandtheimpa

5、ctofisolatedpointsThethesisalsointroducesavarietyofCorpusdepositedinthedictionarystructureaswellastheILseofthestructureofthedictionary；Onthecharacteristicsoftheword，theUseofmultistoreyhashstorage，、vitlIthelargestmatching

6、forwardandachievearapidsegmentationalgorithm，with1GRAM，thespeedofwordsegmentationis2MegabytespersecondThrou曲thecombinationofHiddenMarkovModelPOStaggingandthesmoothingalgorithm，weobtainataggingprecisionof86％，andadisambigu

7、ationof82％InKNNbasedclassificationalgorithm，withtheuseofstatisticalmethodsCHItoselectfeaturewords，andbyloadingthedocumentofrelevantcategorybehindthesewords，wehavesolvedtheproblemofredundantinformationThroughtheUSeofthech

8、aracteristicsofthesentenceandtermweighting，wehaveimplementedthestatisticalmechanicalautomatictextabstractionThroughtheuScofvectorspacemodel，combinedwithsynonymoustermexpansionandinvertedfilestoragestructure，wehaveimpleme