Prediction Of Exchange Rate By Neural Networks

As an important economic variable of financial market,exchange rate has an important effect on macro economy.After the breakdown of the Bretton Woods System,floating exchange rate system has become the main exchange rate system in the west countries. Exchange rate risk beearne one of the most important financial risks,which greatly affects international trade,invest,and finance.Therefore it has been given much attention and all countries and many companies have to take many measures in order to make the best of exchange rate fluctuation.Neural networks have good approximate capability and have been successfully used for prediction.Recently,varieties of neural network models have been used for prediction of exchange rate.The thesis is mainly divided into four chapters:The first chapter gives a primarily introduction of the exchange rate.It briefly explains the meaning of exchange rate,some current exchange rate theory,and analysis factors for the exchange rate.Then,it outlines the basic method of exchange rate prediction and summarizes the state of art of the exchange rate prediction.The second chapter gives a brief introduction of the neural network,which is used to predict the exchange rate in the thesis.It introduces the concept of neural networks and its research history,analyzes the basic principles and the structure of neural networks,and then summarizes the types and the features of neural networks.Finally,it outlines the current trend of the exchange prediction by using the neural network models.The third chapter expatiates the theory,development and algorithm of the Extreme Learning Machine(ELM).Based on those,the fourth chapter establishes the ELM neural network model,and uses it for the prediction of exchange rate.Numerical examples show that ELM network is a good tool for the prediction of exchange rate due to its fast learning rate and high precision.