Pruning for interpretability of large spanned eTS

Abstract

On-line implementation of mechanisms for merging membership functions and rule base simplification are studied in order to improve the interpretability of the eTS fuzzy models. This allows the minimization of redundancy and complexity of the models that may arrive during its development, increasing transparency (human interpretability). The on-line learning technique used is the evolving first-order Takagi-Sugeno (eTS) fuzzy models with rule spanned. A four rule fuzzy system is obtained for the Auto-Mpg benchmark data set with acceptable accuracy