Application Continuous Time Markov Process to Forecast Exchange Rate

In order to reduce error following the improper model chosen to forecast the exchange rate by means of the traditional statistics, the continuous time Markov process is applied to forecast the short time exchange rate, which can describe the frequently fluctuation of exchange rate accurately. Time interval between two state transitions is regarded as a stochastic variable. By the aid of the transition rate matrix, the model is established, and it is solved by the Laplace transform. This proposed method is easy to collect data and calculate the result, and it is effective to detect the state transition. Example shows that when the model is applied to forecast the short-time exchange rate, the forecasted exchange rates have a good agreement with the actual values.