Time series forecasting model using a hybrid ARIMA and neural network

Autoregressive integrated moving average (ARIMA) has been highly popularized, widely used and successfully applied in economic time series forecasting as the popular linear models. Recent research activities in forecasting have focused on artificial neural networks (ANNs) in its generality and practical ease in implementation owing to its powerful and flexible capability. Most studies just use the ARIMA models as the benchmark to test the effectiveness of the ANN model with mixed results. In this paper, a hybrid methodology that combines both ARIMA and ANN models is proposed to take advantage of the unique strength of ARIMA and ANN models in linear and nonlinear modeling. Experimental results with two real data sets (the Wolfs sunspot data and the IBM stock price data) indicate that the combined model can be a more effective way to improve forecasting accuracy than that achieved by either of the models used separately.