Progressive sampling schemes for approximate clustering in very large data sets

摘要: The extensible fast fuzzy c-means algorithm (eFFCM) finds clusters in very large digital images. eFFCM identifies a representative subsample of the image, which is then clustered using (FCM) algorithm. solution extended to secure an approximate clustering remaining pixels image. This article discusses generalized (geFFCM), extension general non-image data. Our accelerates literal (LFCM) on all (loadable) data sets. Second, geFFCM provides feasibility - way find (approximate) for sets that are too be loaded single computer. experiments suggest chi-squared or divergence test goodness fit alone good subsamples. new subsampling method should equally effective acceleration and with VL by any (not just FCM).