Mathematical Formulation of a Type of Hierarchical Neurofuzzy System

摘要

This paper presents a class of hierarchical fuzzy system applied to a cigar classification system. The weight, texture and chromatic characteristics are used to classify multiple cigars using a previous classification based on heuristic knowledge. In the adaptive part, a gradient descendent and error backpropagation method is applied for adjusting the parameters. A detailed description of the algorithm is addressed.