资源描述:
《an outline of data mining methods》由会员上传分享,免费在线阅读,更多相关内容在工程资料-天天文库。
1、6AnoutlineofdataminingmethodsThischapterintroducesthefivechapterswhichformthecoretechnicalcontentofthisbook.Theyarerathermoreaccessiblethansomespecialistbooksonstatistics,dataanalysisandneuralnetworks,andIhopethattheywillbeenjoyabletoread.However,areade
2、rwhoisonlyinterestedintheapplicationsofdataminingandtheproceduresforimplementingitinabusinessmayomitthesechapters.Ontheotherhand,theyareessentialforanyonewishingnotonlytounderstandtheworkingofthetools,inordertousethemmoresuccessfully,butalsotoknowwhena
3、ndwheretouseanyparticularalgorithm.Inthisfirsttechnicalchapter,Ishalloutlinethedescriptiveandpredictivemethodsofdataminingandstatisticsasawhole,andcomparetheirmainfeatures,whichwillbediscussedindetailinthefollowingchapters.Itisimportanttonotethattheloga
4、rithmsusedinthisbookareNapierian(natural)logarithmsinallcases.6.1ClassificationofthemethodsAsmentionedinChapter1,themaindatamininganddataanalysismethodscanbedividedintotwolargefamilies:descriptivemethodsandpredictivemethods.Indescriptivemethods,forreduc
5、ing,summarizingandgroupingdata,thereisnodependentvariable,i.e.noprivilegedvariable.Inpredictivemethods,whichexplaindata,thereisadependentvariable,inotherwordsavariabletobeexplained,oraprivilegedvariable.AmoredetailedversionofthisclassificationisshowninT
6、able6.1,wheremethodsformingpartofconventionalstatisticsanddataanalysishavebeengivengreybackgrounds.Consideringpredictivemethodsonly(Table6.2),wecanbemoreprecisebydistinguish-ingthedifferencesrelatingtothetypeofvariable,namelyindependent(intherows)andde
7、pendent(inthecolumns).Clearly,therows‘nquantitative(representingdifferentquantities)’and‘nqualitative’areonlyrelevantifthedependentvariablesarecorrelatedwitheachother.Otherwise,itissufficienttocarryoutnanalysesofthe‘1quantitative’or‘1qualitative’type.Da
8、taMiningandStatisticsforDecisionMaking,FirstEdition.Ste´phaneTuffe´ry.Ó2011JohnWiley&Sons,Ltd.Published2011byJohnWiley&Sons,Ltd.168ANOUTLINEOFDATAMININGMETHODSTable6.1Classificationofmethods.TypeFamilySub-familyAlgorithmdescriptivegeomet