Modeling and Prediction of the MXNUSD Exchange Rate Using Interval Singleton Type-2 Fuzzy Logic Systems

This paper presents an application of the interval singleton type-2 fuzzy logic system (FLS) to one step ahead prediction of the daily exchange rate between Mexican Peso and US Dollar (MXNUSD) using recursive least-squared (RLS) -back-propagation (BP) hybrid learning. Experiments show that the exchange rate is predictable and according to a simple short-term investment strategy, a good annual profit rate can be obtained. A singleton type-1 FLS and an interval singleton type-2 FLS, both using only BP learning method, are used as a benchmarking systems to compare the results of the hybrid RLS-BP interval singleton type-2 FLS forecaster. The interval singleton type-2 FLS using hybrid RLS-BP learning presents a better performance than both singleton type-1 FLS and interval singleton type-2 FLS.