سال انتشار: ۱۳۸۲

محل انتشار: پنجمین کنفرانس سراسری سیستم های هوشمند

تعداد صفحات: ۹

نویسنده(ها):

Alireza Malekzadeh Arasteh – Senior Student of Electrical Engineering 1 Ferdowsi School of Engineering Mashhad, IRAN

چکیده:

A novel framework termed Complex-valued Fuzzy Logic, which was introduced by D.Ramot, is considered from point of abstract mathematics. Its axiomatic properties were considered briefly and Complex-valued Fuzzy Inference System is presented. The basic concept of Complex-valued Neural Networks is considered to facilitate the way of understanding CANFIS structure. Then, CANFIS, structure as new adoptive network that utilizes both aforementioned structures is introduced. The universal approximator property of CANFIS for real and complex functions is also proved. Finally, the implementation of a sample CANFIS is presented.