資源描述：
《trend and seasonal - regression趨勢和季節(jié)性回歸》由會員上傳分享，免費在線閱讀，更多相關(guān)內(nèi)容在應用文檔-天天文庫。
1、RegressionTechniqueWithDummyVariablesforTrendModelsWithSeasonalVariationSituation:Modelhasbeenobservedtohavelineartrendand4-periodseasonalvariationover20periods.Required:Forecastforthenext4periods(21-24)usingregression.1.Addn-1dummyvariablesforthen-seasons.Sinceinthismodelthereis4-periodseasonal
2、variationadd3dummyvariables.Themodelisnowy=b0+B1t+b2S1+b3S2+b4S3+ε,whereS1,S2,andS3correspondtoseasons1,2,and3respectively.2.HighlightC2:D5andCOPYPutcursorinC6:PASTEPutcursorinC10:PASTEPutcursorinC14:PASTEPutcursorinC18:PASTEPutcursorinC22:PASTEPutcursorinC6:PASTE1.EnterdummyvariablevaluesinC2:D
3、5FordatafromSeason1:EnterS1=1S2=0S3=0.FordatafromSeason2:EnterS1=0S2=1S3=0.FordatafromSeason3:EnterS1=0S2=0S3=1.FordatafromSeason4:EnterS1=0S2=0S3=0.2.MoveColumnBinfrontofcolumnA(ifnecessary)sothatPeriodandSeasonscolumnsarenexttoeachother.3.HighlightB1:B25Select:CUT4.RightMouseClickoncellA1Selec
4、t:INSERTCUTCELLS3.PerformRegressiononValuesvs.Period,S1,S2andS3.5.FillinRegressiondialogueboxRESULT4.Doforecast.6.IncellF22enter:=$H$17+MMULT(B22:E22,$H$18:$H$21)7.DragF22toF23:F25NOTES:1.Inthiscase,MMULT(row,column)isequivalenttoSUMPRODUCT.SUMPRODUCT,however,requiresbothentriesberowsorbothentri
5、esbecolumns.2.Checktoseeifallp-valuesfortheperiodandseasonsarelow.Ifnot,amodifiedF-testcanbeperformedtodetermineifseasonalvariablesshouldbeincluded.
