基于傳遞函數(shù)模型和GARCH模型的金融實證分析.pdf

1、東南大學(xué)碩士學(xué)位論文基于傳遞函數(shù)模型和GARCH模型的金融實證分析姓名：莊玉明申請學(xué)位級別：碩士專業(yè)：概率統(tǒng)計指導(dǎo)教師：陳平20051201AbstractForthecaseoftimeserieswithexogenousvariables，itisbettertosetuptransmissibilityfunctionmodel(TFM)TFMshavebeenwidelyappliedtothefieldsofeconomy、

2、engineering，butisnotsocommonfinance、industrialproductionandinthefieldofstockInthispaper，theTFMisusedtofitandforecastShanghaicompositeindex，whichgetssatisfyingresults。Inordertocomparethefittingandforecastingerrorsofdiffer

3、entmodels，anaveragemethodofmanyperiods’andmultiplesteps’fittingandforecastingerrorsisusedinthispaperThismethodmaydecreasethechanceoffittingerrorandforecastingerrorsoastoincreasethecredenceoftheresultsComparedwiththefitti

4、ngandforecastingerrorsofothermodels，TFMshavebetterresultsthantraditionalARMAmodelsandGARCHmodelsForthetimeserieswithhighlyinterrelatedexogenousvariables，wesuggestanewmethodwithweightedexogenousvariablesThemodelssetupbyth

5、ismethodareadvantageousthanthemodelswithallexogenousvariablesorthemodelskeepingonlyoneexogenousvariablesThecaseinpaperverifiesthisconclusionInthispaper，anewmodeltransmissibilityfunctionmodelwithconditionalhctcrosccdastic

6、ityisproposed。anditsconditionsaswellasthefittingandforecastingeffecthavebeenstudied。WefindthismodelsatisfactoryespeciallyforthetimeserieswithsignificantheterosccdasticityKeywords：timeseries；transmissibilityfunctionmodel；