Using models (P) ARCH and (p,q) GARCH With Daily Data Application for Children Patients

Abstract

The study of the time series includes stationary. The time series models may be stationary or non- stationary, so we study how to deal with the non-stationary time series. The time series be stationary if it was in a statistical equilibrium case and this means that its characteristics don't influence by time .While it be non- stationary if it doesn't include a stable mean and unstable variance.This research includes the analysis of time series representing daily access to patients in Ibn AL-Atheer Teaching Hospital by using the linear ARIMA and non-linear model GARCH (P, q) and how to treat the unstable time series for both types, with empazising stability of the conditional variances predition for GARCH model, and showing to be near to the unconditional variance for GARH model. And choosing models to present these series including linear model ARIMA (4, 1, 3) and non- linear model GARH (1,1). by using (MSE, MAE, MAPE) criteria, of accuracy check to choose the best model from the selected models, by choosing GARCH (1, 1) as the best model.