|
Abstract:
|
Statistical models can provide an explicit framework to develop an insight into infectious disease transmission and prevalence dynamics. Understanding disease patterns helps with identifying risk factors for disease and optimal treatment for clinical practice as well as determining model parameters and prediction of outbreaks. When dealing with contagious diseases, there exist different type of dependencies among patients as well as interactions among multiple risk factors. Therefore modeling consecutive observations of disease patterns in the population over time may be insightful in predicting the association between health outcomes and risks factors. It is of interest to see how the time series models can be applied, evaluated, and extended to explain epidemiological data. In this paper, major classes of time series models such as Auto-Regressive Integrated Moving Average (ARIMA) and Generalized Auto-Regressive Conditional Heteroskedastic (GARCH) have been applied to infectious disease data sets, including an influenza data set. After discovering the best fit, the model is evaluated through comparison of the fit of the time-series methods to the b-spline and wavelet fit.
|
