Exchange rate prediction model analysis based on improved artificial neural network algorithm

Predicting stock prices and currency exchange rates is attracting a great amount of research efforts because of the increasing interests in the prediction models attributed to varying stock markets on a daily basis. This paper investigates a prediction model combined with an ARIMA (Auto regressive integrated moving average model) and a three layer artificial neural network. The complete dataset of from 2010-2013 has been collected, and nine descriptors have been used to train the neural network. The experiment has been tested on the USD/EURO exchange rates. The performance measure is quantified in terms of mean absolute error, mean square error and root mean square error. Experimental results and comparisons demonstrate that the proposed method outperforms the global modeling techniques in terms of profit returns. The predictive power is also clearly shown with a predictor accurately fitting the actual exchange rate data.