Prediction Based On The Exchange Rate Of The Gray Relational Grade Combination

Study on foreign exchange risk becomes more and more paid attention and widely carried out in current economic activities. The basis of foreign exchange management is how to accurately predict the change trend and gradient of the exchange rate. It has important significance and fascinating application value to study on the internal characteristics and prediction of foreign exchange rate, because it is regarded as an important symbol of the international finance market and leverage the balance of the country economy as well.In this paper, based on the analysis of current international and domestic study achievements on prediction technique of exchange rate and the methods combination prediction, a new prediction method of exchange rate based combining prediction theory and grey theory is established and applied to reality.The principle and model of neural network are discussed at first and the focus is the model and algorithm of Grey Theory. Secondly, each single prediction method and several typical linear combining prediction methods are analyzed respectively in detail. Based on the analysis mentioned above and considering the special characteristics of exchange rate, 4 single prediction methods most suitable to the prediction of exchange rate are filtered by fuzzy comprehensive evaluation. Combining these methods and Grey theory , combining prediction model of exchange rate based is build up and applied to the prediction of USD exchange rate. The prediction results are in good agreement with the actual USD exchange rate, which demonstrates that exchange rate can be predicted by the model obtained in this paper efficiently and the exchange rate prediction is verified satisfactorily.