Volume no :2, Issue no: 1, September (2009)

Journal of Statistics: Advances in Theory and Applications

A SIMULATION STUDY OF THE COVARIANCE AND CORRELATION PROPERTIES OF THE FRACTIONALLY INTEGRATED SEPARABLE SPATIAL AUTOREGRESSIVE (FISSAR) MODEL

Author's: Mahendran Shitan and Teng Mei Tuan
Pages: [93] - [111]
Received Date: September 14, 2009
Submitted by:

Abstract

Spatial modelling has its applications in many fields. In time series, there exists a class of models known as long memory models, where the autocorrelation function decays rather slowly. These types of time series data are modelled as fractionally integrated ARMA processes. Spatial data may also exhibit a long memory structure and in order to model such structure, a class of models called as the Fractionally Integrated Separable Spatial Autoregressive (FISSAR) Model has been introduced. The objectives of this research are to demonstrate on how to simulate a FISSAR (1, 1) process, to examine the properties of the variance estimator and to illustrate the correlation properties.

Keywords

fractionally integrated process, autoregressive process, separable models, spatial model, variance, covariance, correlation.