12704.加權(quán)邊緣頻率插值密度估計(jì)

1、加權(quán)邊緣頻率插值密度估計(jì)專業(yè)名稱：統(tǒng)計(jì)學(xué)申請(qǐng)人：張金玲指導(dǎo)老師：楊善朝教授論文答辯委員會(huì)盤燾巫j逸年班愛一芾口E(席員主委WeightedEdgeFrequencyPolygonsDensityEstimatorGrade：2013Major：ProbabilityandStatisticsResearchfield：FinancialStatisticsGraduate：JinlingZhangSupervisor：ShanchaoY

2、angABSTRACTthefrequencyinterpolationestimatederivedfromtheThehistogram，hasbeenwidelyusedInmanydifferentfieldsbecauseofitssimplecalculationmethodResearchonthefrequellcvinterpolationestimated，Scholarshavealwaysbeenmoreconc

3、ernedaboutthisresearchtopicsandpublishedanumberoftheresultsofrelatedworks，Italsobeencontinuousdevelopmentin1998Jones【川eta1，inordertOoptimizethefrequencyinterpolationestimate，proposedaedgefrequencvinterpolationdensityesti

4、mationfEFPcz，=cn6，一1(三七一舌)1[nknk1](三一足jx)lfIntn七。])Wherez∈[(2k—1)k／2，(2k1)6n／／2)，b禮representsthewindowwidthJoneseta1Inthecaseof1ndependentsamplesmeetgivesOptimalwindowwidth、MISEaboutthecorrespondingedgefrequencyhistogram

5、densityestimationEdgefrequencyhistogramdensityestimationonbothsidesofthefrequencvtakenrighttreatment，etcThisarticlewillusetheunequallyweightedmethod，byintroducingaweightingcoefficient入，Weproposedanewclassofweightededgefr

6、equencyinterpolationdensityestimationWEFP(z)=(n％)一1(三七一麥)盼n七(1一A)n七一。](三一七云)[入ntc1一入，n七。])InindependentsamplesSequences，Thispaperresearchrelatedtotheasymptoticnatureofthisnewweightededgefrequencyinterpolationdensityestim

7、ationFirst，studythevarianceitem、biasandmeansquareerroroftheweightededgefrequencyinterpolationdensityestimationunderCOn—ditionsofindependentsamplesAndonthebasisofthat，givestheselectionmethodofoptimalweightandoptimalwindow

8、widthThen，provedtheuniformlystrongconsistencyandasymptoticunbiasednessoftheweightededgefrequencyinterpolationdensityestimationAtlast，Selectthenormaldistributionmodel，andusethethoughtofnumericalsimulation，Thesimulationres