The defuzzification of a type-2 fuzzy set is a two-stage process consisting of firstly type-reduction, and a secondly defuzzification of the resultant type-1 set. All accurate type reduction methods used to build fuzzy classifiers are based on the recursive Karnik-Mendel algorithm, which is troublesome to obtain a feedforward type-2 fuzzy network structure. Moreover, the KM algorithm and its modifications complicate the learning process due to the non-differentiability of the maximum and minimum functions. Therefore, this paper proposes to use the smooth maximum function to develop a new structure of the fuzzy type-2 classifier.
Type-2 Fuzzy Classifier with Smooth Type-Reduction / Nieszporek, K; De Magistris, G; Napoli, C; Starczewski, Jt. - 13588:(2022), pp. 193-202. (Intervento presentato al convegno Artificial Intelligence and Soft Computing. ICAISC 2022. tenutosi a Zakopane, Poland) [10.1007/978-3-031-23492-7_17].
Type-2 Fuzzy Classifier with Smooth Type-Reduction
De Magistris, GSecondo
Formal Analysis
;Napoli, CPenultimo
Supervision
;
2022
Abstract
The defuzzification of a type-2 fuzzy set is a two-stage process consisting of firstly type-reduction, and a secondly defuzzification of the resultant type-1 set. All accurate type reduction methods used to build fuzzy classifiers are based on the recursive Karnik-Mendel algorithm, which is troublesome to obtain a feedforward type-2 fuzzy network structure. Moreover, the KM algorithm and its modifications complicate the learning process due to the non-differentiability of the maximum and minimum functions. Therefore, this paper proposes to use the smooth maximum function to develop a new structure of the fuzzy type-2 classifier.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
