简介:Inthispaper,Usingthedailystockreturndata,weshowthatShanghaistockmarketpricesexhibitlongmemoryprocess,andestimatethelong-memoryparametersbywavelet.Usingthesparsewaveletrepresentationofamatrixoperator,weareabletoapproximateanARFIMAmodelslikelihoodfunctionwiththeseries'swaveletcoefficientsandtheirvariances.MaximizationofthisapproximatelikelihoodfunctionoverthelongmemoryparameterspaceresultsintheapproximatewaveletmaximumlikelihoodestimatesoftheARFIMAmodel.