線性回歸模型回歸系數(shù)的平衡LS估計(jì)及預(yù)測(cè).pdf

1、學(xué)校代號(hào)分類號(hào)10532@淘確火摯HUN刪刪IVERS『rY碩士學(xué)位論文學(xué)號(hào)密級(jí)$03063009線性回歸模型回歸系數(shù)的平衡LS估計(jì)及預(yù)測(cè)學(xué)位申請(qǐng)人姓名趨趨培養(yǎng)單位鏨堂皇鹽量絲渣堂醫(yī)導(dǎo)師姓名及職稱矍遺數(shù)援——學(xué)科專業(yè)褪奎盜里墊堡統(tǒng)讓研究方向紅鹽坌塹皇量選題型論文提交日期2QQ生4旦12日AbstraetInspiredbytheideaofZeller’Sbalancedlossfunction，anewstandardispresen

2、tedforestimatingtheregressioncoefficientoflinearmodelandtheestimationnamedBalancedLSEstimationisgainedunderthenewstandardInChapter3。theexistedtheoriesandmethodsoflinearmodelarediscussedInChapter2，theBalancedLSEstimationi

3、sdefinedforthelinearmodelwhenitserrorcovarianceispositivedefiniteorsingularThenewestimationaswellasthenecessaryandsufficientconditionsisgainedforitsunbiasedness，efficiencyandadmissibilityBecausetheestimationcontainsparam

4、eterstandL，whatthevalueof，andLbringstotheBalancedLSEstimationisdiscussedAnecessaryandsufficientconditionfortheunbiasednessofBalancedLSEstimationisgained，andforthegivenL，tcanbechosentomaketheMSEMoftheBalancedLSEstimationl

5、essthanthatofOLSEstimationForthegivenfwhentheBalancedLSEstimationisunbiased，LcanbechosentomaketheMSEMoftheBalancedLSEstimation1essthanthatofOLSEstimationWhenXiSsatisfiedwithsomeconditions，thesufficientconditionsaregained

6、fortheBalancedLSEstimationtobeemcientthan0LSEstimationInChapter3，theestimationsoftheregressioncoefficientsarestudiedwhentheyarerestrictedbylinearequationsandwhentheyareboundedInChapter4，anewconcept一一balancedLSpredictioni

7、spresentedinthefieldofpredictionofthelinearmodel，andtheoptimalunbiasedbalancedLSpredictionisgainedInChapter5，alinearregressionmodelforthetradedatasissetupWhenthemultilinearityexistsinthedesignmatrixX，theridgeestimationan

8、dtheBalancedLSEstimationisrespectivelyusedtoimproveOLSestimationUnderthiscircumstancestheerrorsquaresofBalancedLSEstimationislessthanthatoftheridgeestimationKeywords：LinearModel；ParameterEstimation；OLSEstimation；Balanced