Construcción de Kernels y funciones de densidad de probabilidad

Abstract

This paper describes  the  details  of  the  well  known  kernel  method  used to construct  a  probability  density  function  for  aunivariate sample. New forms for the kernels are proposed and additional criteria for their election and for estimating the optimal bandwidth. As support for this research a software was implemented for experimentation and obtaining of graphical and numericalresults.