資源描述:
《基于語(yǔ)義的決策樹(shù)挖掘算法研究》由會(huì)員上傳分享,免費(fèi)在線閱讀,更多相關(guān)內(nèi)容在教育資源-天天文庫(kù)。
1、中國(guó)石油大學(xué)(華東)碩士學(xué)位論文基于語(yǔ)義的決策樹(shù)挖掘算法研究姓名:褚希申請(qǐng)學(xué)位級(jí)別:碩士專(zhuān)業(yè):計(jì)算機(jī)應(yīng)用技術(shù)指導(dǎo)教師:時(shí)念云20080501摘要傳統(tǒng)決策樹(shù)算法通過(guò)計(jì)算屬性的信息熵來(lái)選擇屬性,信息熵大的屬性被優(yōu)先選取構(gòu)造決策樹(shù)。在計(jì)算信息熵時(shí),它僅考慮語(yǔ)法層面上字、詞的簡(jiǎn)單匹配,沒(méi)有考慮數(shù)據(jù)的語(yǔ)義信息,缺乏對(duì)其所包含語(yǔ)義信息的理解,這就導(dǎo)致算法缺乏一定的智能性,致使計(jì)算工作量大、復(fù)雜性強(qiáng),而且分類(lèi)質(zhì)量不高。特別是在大數(shù)據(jù)庫(kù)的應(yīng)用上,傳統(tǒng)的決策樹(shù)算法更加面臨大數(shù)據(jù)量計(jì)算的挑戰(zhàn)。本文在分析研究決策樹(shù)挖掘算法及知網(wǎng)、概念樹(shù)、語(yǔ)義相似度等相關(guān)知識(shí)的基礎(chǔ)上,針對(duì)傳統(tǒng)決策樹(shù)挖掘算法的
2、不足,提出了基于語(yǔ)義的決策樹(shù)挖掘思想,實(shí)現(xiàn)了連續(xù)屬性語(yǔ)義化和名詞型屬性語(yǔ)義化的方法,建立了基于語(yǔ)義的決策樹(shù)挖掘模型?;谡Z(yǔ)義的決策樹(shù)挖掘模型較好的利用YN練數(shù)據(jù)中屬性的語(yǔ)義信息,滿(mǎn)足用戶(hù)基于語(yǔ)義的決策樹(shù)挖掘的需求,實(shí)現(xiàn)了一定程度的智能挖掘。實(shí)驗(yàn)表明基于語(yǔ)義的決策樹(shù)挖掘模型能夠解決傳統(tǒng)決策樹(shù)挖掘缺乏語(yǔ)義信息的問(wèn)題、提高數(shù)據(jù)挖掘系統(tǒng)的知識(shí)表示能力,較之傳統(tǒng)的決策樹(shù)挖掘具有更高的效率和預(yù)測(cè)準(zhǔn)確率。關(guān)鍵詞:數(shù)據(jù)挖掘,決策樹(shù),概念樹(shù),語(yǔ)義,智能ResearchoftheDecisionTreeinDataMiningBasedonSemantemeChuXi(ComputerAp
3、plicationTechnology)DirectedbyAssociateProf.ShiNianyunAbstract硼btraditionaldecisiontreealgorithmtakesinformationgainastheruletochoosetheattributeforclassification.,111eattributethathasthebiggestvalueofinformationgaincallbeselectedfirstlytobuildthedecisiontree.Whilecalculatingthevalueofin
4、formationgai瑪thetraditionaldecisiontreealgorithmdoesnotincludesemanticinformation,itonlysimplyconsidersthewords’andcharacters’matchingingrammar,italsolacksoftheunderstandingofthosesernantemeinformationcontainedint11edata.Alloftheaboveresultinlackofintelligenceandleadtoheavycalculation,th
5、ecomplexityandthelow-qualityofclassificationandSOon.Furthermore,thetraditionaldecisiontreealgorithmwillfaceamorebigchallengeforthelargedatabase.Based011theanalysisofthedecisiontreealgorithmandthecorrespondingconceptsuchasHowNet,hierarchytree,sernantemesimilarityandSOon,thispaperproposesa
6、newdecisiontreealgorithmbasedonsernanteme.Thenewalgorithmpresentsthemethodofseparationofcontinuous—attributesandsemantizationofsubstantival-attributes,andbuildsupthesemanteme—baseddecisiontreemodelfordatamining.Thesemanteme-baseddecisiontreemodelCanbetterusethesemantemeinformationaboutth
7、eattributesindatasets;itcallalsomeettheusels’needofdataminingbasedonthesemanteme.Toacertainextent,thesemanteme-baseddecisiontreemodelCanachieveintelligentdatamining.硼1eresultsoftheexperimentsshowthatthesemanteme.baseddecisiontreemodelCannotonlysolvetheproblemoflackingofth