| With the development of economic globalization, the exchange rate used in for-eign currency settlement has become the criterion of commodity circulation in the world. Exchange rate volatility and liquidity of foreign exchange market promote the stability of the exchange rate and safeguard the order of international trade. How-ever, the factors that make the exchange rate change are diversity and mixed together complexly. It makes exchange rate forecasting be a complex nonlinear multivariable system. This interesting and challenging research, attracted wide attention of many scholars both at home and abroad.In recent years, neural networks is increasingly adopted in the prediction of ex-change rate for its nonlinear function approximation ability, strong ability of learning and adaptive ability etc. However, most of them predict a specific number, which can not help the speculators too much who promote the stability of exchange rate because small gap between the predicted values and the actual values will lead to disastrous consequences.In our study, our purpose is to forecast the fluctuation range of the exchange rate. Compared to predict specific values, forecasting the fluctuation range could provide some practical operation reference which is better for speculators than specific values’. In this paper, we present a model named CDBN-FG (Continuous-valued Deep Belief Networks based Fuzzy Granulation) by combining FIG (Fuzzy Information Granula-tion) and DBN (Deep Belief Networks) to forecast the fluctuation range of the ex-change rate. In addition, the concept of "Stop Loss" is introduced by the trading strategy in this paper for making the environment of our profit strategy close to the real foreign exchange trade market. The proposed model is applied to forecasting both Euro/US dollar and British pound/US dollar exchange rate in our experiments. Exper-imental results show that the proposed method is more profitable in the trading process than other typical models. |
PDF Full Text Request