Efficient Forecasting of Exchange rates with Recurrent FLANN

Abstract

In this paper, Functional link artificial recurrent neural network (FLARNN) ensemble forecasting model is proposed for foreign exchange rates prediction. In the survey of existing literature, it is revealed that there is need to develop efficient forecasting models involving less computational load and fast forecasting capability. We compare the FLARNN model performance with some existing neural network models such as LMS and FLANN in terms of the exchange rates between US Dollar and other three currencies (Indian Rupees, British pound, Japanese Yen). Experimental results reveal that the predictions using the proposed approach are consistently better than the above mentioned methods on the basis of various parameters including error convergence and the Mean Average Percentage Error (MAPE).