An artificial neural network framework for dual interest rate parity

In an effort to improve the forecasting accuracy, the one way arbitrage relationship that links the forward market with the international money market, is extended to include the currency futures market. The dual interest parity relationship formulated is then used to design an artificial neural network model that forecasts future spot rates of exchange. Exchange rate forecasts generated by the network, when compared to those predicted by the forward and futures rates, were found to be more accurate and better at predicting the directional movement of currency prices.