The Effects of Parameter on Power Spectral Density function for Moving Average Model from First Order: A Simulation Study

Abstract

The present paper investigates the effects of an estimated parameter by Iterative Process for Non-Gaussian Moving Average Model from first order on power spectral density function ( or PSD) for samples of different sizes and parameter’s initial values of different values and signs, regarding that random errors distribute either discrete or continuous. This is done by a simulation study. Among the important findings are (1) PSD values increase when the parameter sign is (-) and decrease when it is ( +) for all discrete and continuous distributions; (2) PSD values increase when the sample size is deceasing ; and (3) PSD values in discrete distribution is less than those in continuous distribution.