基于粗糙集的多变量决策树在电力市场营销分析中的应用

《基于粗糙集的多变量决策树在电力市场营销分析中的应用》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、ABSTRACTWiththerapiddevelopmentofdomesticelectricpowerindustryandthedeep-goingofthemarket-orientedsystemreformprocessofit,thedutyofthepowercompanieshasbeencringingfromthecustomermanagementtothemarketingandservice．Therefore，electricmarketinganalysis，whichdec

2、ideswhetherthepowercompaniescansurviveundertheconditionofthedl'asticmarketcompetitionandgetdevelopment,becomesthegroundworkofthepowercompanies．However,mosttraditionalmethodsforelectricmarketinganalysisarefocusondatamanagementorjustdosomesimpledatastatistica

3、nalysis．ItisverydifficulttOgainsignificantinformationformarketingdecisionsjustbythesemethods，whenfacingmoreandmoreenorID_01．ISdatabaseofelectricmarketing．Aimingatupperquestions，inthispaperweapplythemultivariatedecisiontreebasedonroughsettoelectricmarketinga

4、nalysis．Firstly,segmenttheelectricmarketaccordingtothedifferenceofindustryandmonth,ete．Thenapplythedecision慨whichisbasedOilroughsettos呻therelationshipbetweenthemarketingbehaviorandtheinfluencingfactors,suchasprice，precipitation,temperatureete，andquantifythi

5、srelationshipwithcertainmeasurementindexes．AccordingtOtheresults，themarketinganalystcallgetmoreveraciousjudgementandunderstandingOilthecharacteristicsofpowermarket．andn妇correspondingmarketingstrategies．Basedontheideas，alotofpracticalworkhasbeendone．Firstly,

6、datawarehousewiththethemeofelectricmarketinganalysisisbuilt．ThentheclusteringmethodofanimprovedK-Meansisappliedtogeneralizetheoriginaldata．Secondly,usingroughsettheorytOreducethenumberofattributesandkeeponlytheessentialattributes，andamultivariatedecisiontre

7、eisconstructed．Finally,Onthebasisofthedecisiontree，therelationshipbetweenelectricmarketingandtheinfluencingfactorscanbeeasilyobtained，meanwhileWetestifytheresultstoeliminatethepossiblemisguidanceinthemandguaranteetheveracityoftherules．Accordingtothepractica

8、lexample,itshowsthatthemethodgiveninthispapercallacquiretherulesinelectricpowermarketingfullyandquickly,andtherulesmaynotbefoundeasilywithtraditionalmethods．Themultivariatedecisiontreetechniquebasedonr