資源描述:
《基于語義的決策樹挖掘算法研究》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。
1、中國石油大學(xué)(華東)碩士學(xué)位論文基于語義的決策樹挖掘算法研究姓名:褚希申請學(xué)位級別:碩士專業(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)決策樹算法通過計(jì)算屬性的信息熵來選擇屬性,信息熵大的屬性被優(yōu)先選取構(gòu)造決策樹。在計(jì)算信息熵時(shí),它僅考慮語法層面上字、詞的簡單匹配,沒有考慮數(shù)據(jù)的語義信息,缺乏對其所包含語義信息的理解,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),而且分類質(zhì)量不高。特別是在大數(shù)據(jù)庫的應(yīng)用上,傳統(tǒng)的決策樹算法更加面臨大數(shù)據(jù)量計(jì)算的挑戰(zhàn)。本文在分析研究決策樹挖掘算法及知網(wǎng)、概念樹、語義相似度等相關(guān)知識(shí)的基礎(chǔ)上,針對傳統(tǒng)決策樹挖掘算法的
2、不足,提出了基于語義的決策樹挖掘思想,實(shí)現(xiàn)了連續(xù)屬性語義化和名詞型屬性語義化的方法,建立了基于語義的決策樹挖掘模型。基于語義的決策樹挖掘模型較好的利用YN練數(shù)據(jù)中屬性的語義信息,滿足用戶基于語義的決策樹挖掘的需求,實(shí)現(xiàn)了一定程度的智能挖掘。實(shí)驗(yàn)表明基于語義的決策樹挖掘模型能夠解決傳統(tǒng)決策樹挖掘缺乏語義信息的問題、提高數(shù)據(jù)挖掘系統(tǒng)的知識(shí)表示能力,較之傳統(tǒng)的決策樹挖掘具有更高的效率和預(yù)測準(zhǔn)確率。關(guān)鍵詞:數(shù)據(jù)挖掘,決策樹,概念樹,語義,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