Exchange rate forecasting better with hybrid artificial neural networks models

Abstract

Forecasting accuracy is one of the most important features of forecasting models; hence, never has research directed at improving upon the effectiveness of time series models stopped. Nowadays, despite the numerous time series forecasting models proposed in several past decades, it is widely recognized that exchange rates are extremely difficult to forecast. Artificial neural networks (ANNs) are one of the most accurate and widely used forecasting models that have been successfully applied for exchange rate forecasting. In this paper, an improved model of the artificial neural networks is proposed using autoregressive integrated moving average models, in order to yield more general and more accurate hybrid model than artificial neural networks for time series forecasting. In our proposed model, the unique advantages of the ARIMA models in linear modeling are used in order to preprocess the under-study data for using in artificial neural networks. Empirical results in weekly Indian rupee against the United States dollar exchange rate indicate that the proposed model can be an effective way to improve forecasting accuracy achieved by artificial neural networks and traditional linear models.