Radial Basis Function Neural Network for Time Series Prediction: The Case of Yen Currency

In this work, an ensemble radial basis network is developed, this design is used for time series prediction. This method consists of finding the number of neurons and the number of hidden layers and the integration of network responses carried out by Type-1 and Type-2 Fuzy System and are of Mamdani type (FIS), The tests are carried out with Yen/Dollar and Yen/Mexican exchange rates. This demonstrated that the model is successful, as it gives good prediction results.