Divulgação de seminário

["Roseli A. Leandro" <raleandr@carpa.ciagri.usp.br>]

 O PPG em Estatística e   Experimentação Agrônomica, ESALQ, USP,
 convida a todos  para assistir o seminário:

 *** Discrete kernel estimators for a probability mass function ***
                         **Célestin C. Kokonendji**
   Université de Pau et des Pays de l´Adour,  France

 Local:  Departamento de Ciências Exatas, ESALQ/USP, Piracicaba, SP
        Sala 311******  Data: 16/03/2006 ********   Horário: 16:00 h



 *****************************************
 Abstract: The main of this talk is to introduce the discrete kernel
 estimators for estimating a probability mass function (pmf). The
 standard continuous kernels are, of course, unsuitable for fitting
 count data distributions. From an appropriated definition of the
 discrete kernel estimator of a pmf, we point out some basic
 (asymptotical) properties using finite difference approximation for
 the derivation. In particular, systematic and stochastic errors and,
 then, mean (integrated) squared errors are investigated under general
 considerations. The Poisson, binomial, negative binomial and uniform
 discrete kernels are used for illustrations.

 *******************************************

 Cordiais saudações, Roseli.