Accurately forecasting the foreign exchange rate is critical to enterprises like the semiconductor industry in Taiwan. For this purpose, a fuzzy-neural approach is proposed. In the proposed methodology, a committee of virtual experts is organized instead, and then they are asked to give opinions about the fuzzy forecasts. A corresponding FLR equation is constructed to forecast the foreign exchange rate for each virtual expert. To aggregate these fuzzy foreign exchange rate forecasts, a two-step aggregation mechanism is applied. First, partial-consensus fuzzy intersection is applied to aggregate the fuzzy forecasts into a polygon-shaped fuzzy number, in order to improve the precision. Then a radial basis function network (RBF) is constructed to defuzzify the polygon-shaped fuzzy number and to generate a representative/crisp value, so as to enhance the accuracy. To evaluate the effectiveness of the proposed methodology, the practical case of forecasting the foreign exchange rate of NTD for USD is used. According to the experimental results, the proposed methodology improved.
展开▼
