Building a Hybrid Model to Forecast the Numbers of Children with Respiratory Diseases in Kirkuk Governorate

Abstract

Hybrid models that combine linear and nonlinear models are among the most important tools for analyzing time series, This article discussed a methodology for building hybrid time series models and its application in identifying the behavior of the monthly series of the number of children admitted to the Children's Hospital in Kirkuk Governorate related to respiratory diseases.. By applying the linear models to the studied series  then using the residuals of the linear model as inputs for the later stages, we were able to find the best and optimal hybrid model from all the models that were evaluated in this study according to the criteria that were used to determine the best model, the study concluded that the hybrid model (ARIMA-EGARCH) was the optimal model that can be relied upon in making forecasting for the series studied in the application.