AR(1) واألنموذج ARIMA(0,1,0) استخدام المحاكاة للتمييز بين أنموذج المسار العشوائي

Abstract

The objective of This Paper is to distinguish between random walk model and model through the identification of the cross – over point for autocorrelation function and the properties of this function by using simulation method. And have been obtained of closing estimates for real behavior from the two models above by assuming that presenting random error series which is distributed as uncorrelated normal distribution. With zero mean and variance is equal (1) without losing the generality. Three simulation experiments with sample size (200, 100, 50) and the number of replicates (1000) was carried. The measure of empirical distribution of the correlation, namely, the mean and standard deviation for autocorrelations sample in addition to the skewness and kourtis was computed. Some conclusions and suggestions of the study was carried. Such that, the autocorrelation coefficients of model are changing from positive to negative signal faster than autocorrelations coefficients of random walk model.