Linear and nonlinear univariate models applied to forecasting financial series

This work investigates the hypothesis that the nonlinear models of feedforward and radial basis function neural networks and the Takagi-Sugeno (TS) fuzzy system are able to provide a more accurate forecast than the traditional ARMA and ARMA-GARCH linear models. Using series of Brazilian exchange rate (R$/US$) returns with 15 min., 60 min., 120 min., daily and weekly basis, the forecast performance is compared. Results indicate that forecast performance is related to the series' frequency and the forecasting evaluation shows that nonlinear models perform better than their linear counterparts. In the trade strategy, nonlinear models achieve higher returns when compared to a naive strategy and to the linear models.