人工神經(jīng)網(wǎng)絡(luò)和決策樹(shù)進(jìn)行數(shù)據(jù)分類(lèi)的對(duì)比研究.pdf

1、學(xué)位論文厶三塑塑旦塑塑達(dá)筮塑鱟i重塑堡坌耋塑墮!!墮塞(Ⅱ目)指導(dǎo)教師申請(qǐng)學(xué)位級(jí)別專(zhuān)業(yè)名稱學(xué)位授予單位陣飛作#g史永文副教授工學(xué)碩士計(jì)算機(jī)應(yīng)用技術(shù)太蓮海事大學(xué)2005年3月AbstractWiththematurityofdatabasetechnologyandthepopularizationofdataapplicationusinginhumansocietydataamountthatmankindaccumulateisin

2、creasinginincrediblyrapidsp。edMenawinf弛eoftheendlessoceanofinformationandthehugewaveofthedataisbillowingandwellinguptopeople；o雛theotherhandthesituationthatcontrastedsharplywiththepresentis“Wearcdrowningininformation，buts

3、tarvingforknowledge”P(pán)artofthereasonisduetolackthehlghspeed，highefficientmethodmakingu∞ofinformationtechnologytogetusefulinformationandknowledgePeopleareatatotaltosswhattodofacingthehugeamountofdataInordertosolvethecontra

4、dictionbetweenthedataenrichingandpoorknowledge，theresearchershavedevelopedKnowledgeDiscoveryinDatabase(KDD)anditskeyteclmologp—Dg堍Mining∞扣9DataclassificationisoneofthemostimportantmethodsindataminingHavethen‰eYousanddiso

5、rderlydatapeopleshouldcI∞s瓣thesedatafirst，thenanalysesandstudythelessdatasetsrespectivelyandthenmakingdecisionandpredictionBasedonthebackgroundsmentionedabove，westudytwokindsofpopularandimportantda詛classificationme噸ods_卅

6、ecisiontreeandartificialneuralnetworksWeintroducetheirdefinitionsconstructionmethodandmodelsAccordingtothepracticalconditions，bothofthetwomethodsatecarefullyandcompletelystudiedincontrastfromdifferentaspea：；：tsanddiffere

7、ntlevelsTheconclusionthatthispapergetsfinallyis，underthespecificenvironmentandspecifiedconditions，asakindofnewtechnologytheartificialneuralnetworksmethodhasdistinctivepredominanceinsolvingthedataclassificationproblems，Ke