سال انتشار: ۱۳۸۷

محل انتشار: دومین کنگره مشترک سیستم های فازی و سیستم های هوشمند

تعداد صفحات: ۸

نویسنده(ها):

M Alipoor – Engineering Department of Tarbiat Moallem University of Sabzevar, Sabzevar, Iran
H. Sadoghi Yazdi –

چکیده:

Conventional clustering methods may fail to solve some real world clustering problems. In this paper, a novel clustering method using PSO-adjustedspace conversion is presented. The main idea of proposed algorithm is to transfer given dataset to a new space in witch conventional clustering methodscan determine the intrinsic grouping in the set. Finding such a linear transform that would result in minimum error and precise clustering is performed by particleswarm optimization (PSO). The merit of proposed method is certificated in skin detection as a real word benchmark problem