A Hybrid Arima And Artificial Neutral Network In Forecasting

Abstract

Autoregressive Integrated Moving Average models (ARIMA) is a popular choice of many statisticians or corporate forecasters in financial forecasting during the past three decades. However, the drawback of ARIMA models is that they assume the future values of a time series have a linear relationship with current and past values as well as with white noise, so using ARIMA models may not be suitable for complex nonlinear problems. Then Artificial Neural Network (ANN) models have been developed and become an alternative to overcome the limitation of the linear time series models, nevertheless their performance in some situations is inconsistent. In this study, we propose a hybrid model which is a combination of both ARIMA and ANN to satisfy the real-world time series that is generally contains both linear and nonlinear patterns for improving forecasting efficiency.