简介：Inordertoconstructestimatingfunctionsinsomeparametricmodels,thispaperintroducestwoclassesofinformationmatrices.Somenecessaryandsufficientconditionsfortheinformationmatricesachievingtheirupperboundsaregiven.Fortheproblemofestimatingthemedian,someoptimumestimatingfunctionsbasedontheinformationmatricesareacquired.Undersomeregularityconditions,anapproachtocarryingoutthebestbasisfunctionisintroduced.Innonlinearregressionmodels,anoptimumestimatingfunctionbasedontheinformationmatricesisobtained.Someexamplesaregiventoillustratetheresults.Finally,theconceptofoptimumestimatingfunctionandthemethodsofconstructingoptimumestimatingfunctionaredevelopedinmoregeneralstatisticalmodels.