Time series forecasting using a hybrid ARIMA and neural network model

Experimental results with real data sets indicate that the combined model can be an effective way to improve forecasting accuracy achieved by either of the models used separately.