作者: Abdelhamid Bouchachia , Markus Prossegger
DOI: 10.1007/978-1-4419-8020-5_4
关键词:
摘要: In the present contribution, a novel algorithm for off-line spectral clustering is introduced and an online extension derived in order to deal with sequential data. The proposed aims at dealing nonconvex clusters having different forms. It relies on notion of communicability that allows handle contiguity data distribution. second part paper, incremental fuzzy c-varieties serve as building block (ISC). Initial simulations are presented towards end contribution show performance ISC algorithm.