Robust fuzzy tree model dealing with unstructured bounded data uncertainties

In projects, the foundation of a robust model for measured data matrices with (not necessarily small) deterministic bounded uncertainties is a common problem. Employing robust least squares (RLS) linear modeling resolutions with these perturbation data, in the field of nonlinear models, based on several fuzzy rules, the robust fuzzy tree model (RFT) provides a cheerful method of resolution. This paper discusses a certain RFT model for data with unstructured bounded uncertainties. Furthermore, a performance comparison of simulation examples between the RFT and traditional fuzzy tree model (FT) prove advantageous over the RFT. The RFT not only keeps the ability of dealing with the high dimensional input problem and the features of less computation load and high preciseness of FT, but also decreases drastically the sensitivity of the FT to bounded uncertainties.