离散选择模型logit模型实例stata分析ppt课件.pptx

《离散选择模型logit模型实例stata分析ppt课件.pptx》由会员上传分享，免费在线阅读，更多相关内容在教育资源-天天文库。

1、DiscreteChoiceAnalysisTermpaperLogitmodelingwithstata2016.12.14G201538010LITINGTING1ContentsIntroductionSPDatasetinformationModelingscenariossettingDatamodificationModelingestimatedresultsModelingcomparisonConclusion2IntroductionThispaperdevelopedadisaggregatedlogisticsdemandmodelsusingdiscretech

2、oiceanalysismethod.Datausedis2008-SPdatafromasurvey.Statawasemployedfortheestimationoflogitmodels.3SPDatasetinformationStatedpointed:2007Analysiscommodity:steelAnalysisrange:30people(row600-1500)Dependentvariable:choiceIndependentvariables:cost[log(#/10^5)]time[hour]LOS[#/10]4ModelingScenariosSett

3、ingModelingscenariosMarketSegment?YNMODEL1-1-Distance5/63variables(cost/time/LOS)MODEL2-1-Distance5/62variables(cost/time)MODEL3-1-Distance5/62variables(cost/LOS)MODEL4-1-Distance5/62variables(time/LOS)MODEL1-23variables(cost/time/LOS)MODEL2-22variables(cost/time)MODEL3-22variables(cost/LOS)MODEL4

4、-22variables(time/LOS)5DataModificationWemodifyrow-datatoremoveunreasonabledataset-Suchasthechoiceofthenot-dominantalternativeModel1has1unreasonabledatasets(inalldatasets)Model2has31unreasonabledatasets(inalldatasets)Model3has8unreasonabledatasets(inalldatasets)Model4has85unreasonabledatasets(inal

5、ldatasets)Model4isignored6idmodedistanceqtypecosttimeloschoicelogcost5los2223rail51210000146000.7419376223truck5128000066011.0296196223rail5102100001910000.74193710223truck51028000076011.0296196Datareorganizeexampleiddistancequestiontyperailtruckchoicecosttimeloscosttimelos223512100001460280000660

6、22235226000014803500007602BasicdataDataModification7ModelingEstimatedResults(DIST5)Modeldistance51-1-50.28992-1-50.28843-1-50.10428ModelingEstimatedResults(DIST6)Modeldistance61-1-60.25882-1-60.25393-1-60.07049ModelingEstimatedResults(DIST5&6)ModelDistance5&61-20.25392-20.25213-20.083810ModelingCo

7、mparison11ConclusionThebestmodelisModel1-1Statisticalresultsoftheestimationindicatethatmarketsegmentationmakesthemodelmorecomprehensive.Themorevariablesmodelhas,themorecomprehensivemodelis.Themodelwhichisused3var