Exchange Rate Model Approximation, Forecast and Sensitivity Analysis by Neural Networks, Case Of Iran

Abstract

This paper investigates the model estimation and data forecasting of exchange rate using artificial neural network. Recent studies have shown the classification and prediction power of the neural networks. It has been demonstrated that a neural network can approximate any continuous function. Here, in a technical approach, it has been used ARIMA and neural network for a short-term forecast of daily USD to Rial exchange rate. ANN is employed in training and learning processes and thereafter the forecast performance measured making use of two common loss functions. The comparison demonstrates that neural network is far better than ARIMA, the error is about the half.Thereafter, in a fundamental approach via another neural network the effects of some of the most important economic variables on exchange rate prediction in a long-term sense are studied. By sensitivity analysis, the importance and the weight of each economic variable on exchange rate has determined. The results show that it is possible to estimate a model to forecast the value of exchange rate even by having access to a limited subset of data.

Utility of a Shuffled Differential Evolution algorithm in designing of a Pi-Sigma Neural Network based predictor model

“…The results reveal the better performance of ANN over ARMA for both the trend and deviation prediction. Pedram and Ebrahimi, in [5] have compared the performance of an ANN with an Auto Regressive Integrated Moving Average (ARIMA) model for a short-term forecast of daily USD to Rial exchange rate. The results have clearly shown that the error statistics observed from ANN is nearly half of the ARIMA model.…”