Combining Rule Weight Learning and Rule Selection to Obtain Simpler and More Accurate Linguistic Fuzzy Models

In complex multidimensional problems with a highly nonlinear input-output relation, inconsistent or redundant rules can be found in the fuzzy model rule base, which can result in a loss of accuracy and interpretability. Moreover, the rules could not cooperate in the best possible way. It is known that the use of rule weights as a local tuning of linguistic rules, enables the linguistic fuzzy models to cope with inefficient and/or redundant rules and thereby enhances the robustness, flexibility and system modeling capability. On the other hand, rule selection performs a simplification of the previously identified fuzzy rule base, removing inefficient and/or redundant rules in order to improve the cooperation among them. Since both approaches are not isolated and they have complementary characteristics, they could be combined among them. In this work, we analyze the hybridization of both techniques to obtain simpler and more accurate linguistic fuzzy models.