A Nonexclusive Classification System Based on Co-operative Fuzzy Clustering

Nonexclusive classification characterizes many real problems in which a hard decision about data labels cannot be taken. In these cases, a decision system capable of fuzzy outputs is desirable, in order to well describe problem's nature and domain. In this paper, an algorithm pursuing this approach is presented. More precisely, a nonexclusive k-class problem is solved by the co-operation of k independent clustering systems. In order to better evaluate and compare the presented neuro-fuzzy classifier in simulation tests, we also propose a fuzzy classification quality (FCQ) measure.