Time Series Forecasting Using Hybrid AutoRegressive Integrated Moving Average(ARIMA) and Artificial Neural Network (ANN) Model

Abstract

Autoregressive Integrated Moving Average (ARIMA) is one of the popular linear models in time series forecasting during the past three decades. Recent research activities in forecasting with Artificial Neural Networks (ANN) suggest that ANN can be a promising alternative to the traditional linear methods. ARIMA models and ANN are often compared with mixed conclusions in terms of the superiority in forecasting performance. In this thesis, a hybrid methodology that combines both ARIMA and ANN models is also proposed to take advantage of the unique strength of ARIMA and ANN models in linear and nonlinear modeling. Experimental results with real datasets indicate that the combined model can be an effective way to improve forecasting accuracy achieved by either of the models used separately.