Please use this identifier to cite or link to this item:
|Title: ||Generalized Integer-Valued Random Coefficient for a first Order Structure Autoregressive (RCINAR) Process|
|Authors: ||Gomes, Dulce|
Canto e Castro, Luisa
|Editors: ||DasGupta, A.|
|Keywords: ||INAR models|
stochastic autoregressive models
generalized thinning operation
|Issue Date: ||28-May-2009|
|Publisher: ||Journal of Statistical Planning and Inference|
|Citation: ||Zacarias OP, Majlender P.
Malar J. 2011 Apr 17;10:93. doi:10.1186/1475 2875-10-93.
Spiegelman D, Casella M.
Biometrics. 1997 Jun;53(2):395-409.|
|Abstract: ||A random coefficient autoregressive process for count data based on a generalized thinning operator is presented. Existence
and weak stationarity conditions for these models are established. For the particular case of the (generalized) binomial thinning, it is proved that the necessary and sufficient conditions for weak stationarity are the same as those for continuous valued AR(1) processes. These kinds of processes are appropriate for modelling non-linear integer-valued time series. They allow for
over-dispersion and are appropriate when including covariates. Model parameters estimators are calculated and their properties studied analytically and/or through simulation.|
|Appears in Collections:||MAT - Publicações - Artigos em Revistas Internacionais Com Arbitragem Científica|
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.