简介:Researchonjumpregressionfunc(?)onshasnotbeenadequateyetAccordingtotheinformationaboutthenumberofjumps,theirpositionsandjumpmagnitudes,jumpregressionfunctionscanbeclassifiedintoeighttypes.Thispaperdealsespeciallywiththesecondjumpregressionfunction.Firstofall,aconceptoftrimmedsplineestimateisproposedandwithitanL~2-consistentestimateofthesmoothingpartofthejumpregressionfunctionisobtained.ThisalongwiththeL~2-consistentestimateofjumpmagnitudeconstitutesanestimateofthesecondjumpregressionfunction.Thispaperdiscussesalsothecasethatthejumppositionshavesomeindeterminacy.Anewcriterionissuggestedanditsuniquesolutionderived.Intheend,afewnumericalresultsaregiven.
简介:Inthispaper,thedefinitionsofBayesestimator,minimaxestimator,minimumriskequivariantestimatorandadmissibleestimatorunderthegeneralmatrixlossfunctionaregiven.Severalresultsonestimationofparameterswiththeordinarylossfunctionareextendedtothecaseofthematrixlossfunction.
简介:Stronguniformconsistencyratesaregivenforkerneltypeestimatorsoftheconditionalfunctionwith(?)-mixingsample.Especially,fornonparametricestimatorsofkerneldensity,theregressionfunctionwhenYisbounded,conditionaldf’s,L-smoothingandM-smoothing,weobtainthesamerateO((n/logn)-1/3)asinthei.i.d.sampleestablishedbyH(?)rdle,JanssenandSerfling.
简介:EXTREMEPOINTRESULTSFORSTRICTPOSITIVEREALNESSOFTRANSFERFUNCTIONFAMILIESWANGLong;HUANGLin(DopartmentofMechanics,PekingUniversit...
简介: