Introduction on a forecasting model for infectious disease incidence rate based on radial basis function network / 中华流行病学杂志

ABSTRACT

It is important to forecast incidence rates of infectious disease for the development of a better program on its prevention and control. Since the incidence rate of infectious disease is influenced by multiple factors, and the action mechanisms of these factors are usually unable to be described with accurate mathematical linguistic forms, the radial basis function (RBF) neural network is introduced to solve the nonlinear approximation issues and to predict incidence rates of infectious disease. The forecasting model is constructed under data from hepatitis B monthly incidence rate reports from 1991-2002. After learning and training on the basic concepts of the network, simulation experiments are completed, and then the incidence rates from Jan. 2003-Jun. 2003 forecasted by the established model. Through comparing with the actual incidence rate, the reliability of the model is evaluated. When comparing with ARIMA model, RBF network model seems to be more effective and feasible for predicting the incidence rates of infectious disease, observed in the short term.