A nonlinear forecasts combination method based on Takagi–Sugeno fuzzy systems

In this paper, we investigate the use of a special class of fuzzy systems, namely first order Takagi–Sugeno fuzzy systems, to combine a set of individual forecasts. Such systems can be interpreted as local linear approximation models and have been used mainly as such in this study. The inference produced by these models can be seen as a new kind of piecewise linear regression with softened transitions between the pieces. By comparing our combining system with traditional linear combining models, we have shown the possible advantage of the nonlinear approach as well as the flexibility of our system.