- 期刊
- OpenAccess
Fuzzy Systems Neural Networks and Makov Switching AR Model for Prediction of Exchange Rates
並列摘要
Many international agents (e.g. money managers, investment banks, investors, funds makers and others) are very concerned about predicted values of exchange rates because it often moves dustically and generally affects the profits. This paper forecasted the daily Bangladeshi and Canadian exchange rates for the period of October 1996 to January 2013. With attention paid to recently developed econometric noises, the widely‐used forecasting model the fuzzy extension of artificial neural network is considered and compared results with the Markov switching autoregressive forecasting model. Root mean square error and correlation coefficient are used as an evaluation measures. It has been found that the fuzzy extension of the artificial neural network model is a superior predictor compared to the other selected predictor for the Bangladeshi series and the reverse observed for the Canadian series. It is believed that the findings of this paper will be helpful for multinational organizations wanting to make wise policy about these two country's exchange rates.