Factor analysis latent subspace modeling and robust fuzzy clustering using t-distributions

Chatzis, S; Varvarigou, T

dc.contributor.author	Chatzis, S	en
dc.contributor.author	Varvarigou, T	en
dc.date.accessioned	2014-03-01T01:30:38Z
dc.date.available	2014-03-01T01:30:38Z
dc.date.issued	2009	en
dc.identifier.issn	1063-6706	en
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/19607
dc.subject	Factor analysis	en
dc.subject	Fuzzy clustering	en
dc.subject	Kullback-Leibler divergence	en
dc.subject	Local subspace modeling	en
dc.subject	Student's t-distribution	en
dc.subject.classification	Computer Science, Artificial Intelligence	en
dc.subject.classification	Engineering, Electrical & Electronic	en
dc.subject.other	Clustering approach	en
dc.subject.other	Clustering procedure	en
dc.subject.other	Clustering scheme	en
dc.subject.other	Dimensionality reduction	en
dc.subject.other	Distributed data	en
dc.subject.other	Expectation maximization	en
dc.subject.other	Factor analysis	en
dc.subject.other	Fuzzy C-means	en
dc.subject.other	Gaussian mixture models	en
dc.subject.other	Kullback-Leibler divergence	en
dc.subject.other	Local subspace modeling	en
dc.subject.other	Model yields	en
dc.subject.other	Objective functions	en
dc.subject.other	Observation space	en
dc.subject.other	Robust fuzzy clustering	en
dc.subject.other	Subspace models	en
dc.subject.other	T distribution	en
dc.subject.other	T-mixture models	en
dc.subject.other	Blind source separation	en
dc.subject.other	Cluster analysis	en
dc.subject.other	Clustering algorithms	en
dc.subject.other	Fuzzy rules	en
dc.subject.other	Fuzzy systems	en
dc.subject.other	Students	en
dc.subject.other	Fuzzy clustering	en
dc.title	Factor analysis latent subspace modeling and robust fuzzy clustering using t-distributions	en
heal.type	journalArticle	en
heal.identifier.primary	10.1109/TFUZZ.2008.924317	en
heal.identifier.secondary	http://dx.doi.org/10.1109/TFUZZ.2008.924317	en
heal.language	English	en
heal.publicationDate	2009	en
heal.abstract	Factor analysis is a latent subspace model commonly used for local dimensionality reduction tasks. Fuzzy c-means (FCM) type fuzzy clustering approaches are closely related to Gaussian mixture models (GMMs), and expectation - maximization (EM) like algorithms have been employed in fuzzy clustering with regularized objective functions. Student's t-mixture models (SMMs) have been proposed recently as an alternative to GMMs, resolving their outlier vulnerability problems. In this paper, we propose a novel FCM-type fuzzy clustering scheme providing two significant benefits when compared with the existing approaches. First, it provides a well-established observation space dimensionality reduction framework for fuzzy clustering algorithms based on factor analysis, allowing concurrent performance of fuzzy clustering and, within each cluster, local dimensionality reduction. Second, it exploits the outlier tolerance advantages of SMMs to provide a novel, soundly founded, nonheuristic, robust fuzzy clustering framework by introducing the effective means to incorporate the explicit assumption about Student's t-distributed data into the fuzzy clustering procedure. This way, the proposed model yields a significant performance increase for the fuzzy clustering algorithm, as we experimentally demonstrate. © 2009 IEEE.	en
heal.publisher	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC	en
heal.journalName	IEEE Transactions on Fuzzy Systems	en
dc.identifier.doi	10.1109/TFUZZ.2008.924317	en
dc.identifier.isi	ISI:000266677000002	en
dc.identifier.volume	17	en
dc.identifier.issue	3	en
dc.identifier.spage	505	en
dc.identifier.epage	517	en