Analyzing the Maximum Lyapunov Exponent by Chaos Theory in Iran Foreign Exchange Rate

Nonlinear dynamic systems exhibit different behaviors, that can be utilized to explain many economic phenomena that seems to be stochastic. Chaos theory suggests a new method to study the changes of nonlinear dynamic systems in the financial markets. In this paper, we studied Iran foreign exchange rate sensitivity to initial conditions by chaos theory and maximum Lyapunov exponent against the U.S. and Canadian dollar, British Pound, the Euro and the UAE Dirham during 24/03/1982 to 23/05/2007. For this purpose, we analyze the presence of chaos behavior in the above mentional currencies by strong correlation test and maximum Lyapunov exponent. The obtional result indicate that Iran foreign exchange rate against the U.S. dollar is and fellows a chaos process. Therefore the linear methods are not appropriate for prediction of the variable. In second part of this paper, we predict the Iran exchange rate against the U.S. dollar by nonlinear neural Network model using optimizing self comparative particles group algorithm. Results of neural network algorithm shows that daily exchange prices in a short period are highly predictable based on prior price.