Vertical collaborative fuzzy C-means for multiple EEG data sets

Publication Type:
Conference Proceeding
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2013, 8102 LNAI (PART 1), pp. 246 - 257
Issue Date:
Filename Description Size
10.1007%2F978-3-642-40852-6_26.pdfPublished version280.32 kB
Adobe PDF
Full metadata record
Vertical Collaborative Fuzzy C-Means (VC-FCM) is a clustering method that performs clustering on a data set of having some set of patterns with the collaboration of some knowledge which is obtained from other data set having the same number of features but different set of patterns. Uncertain relationship lies in data between the data sets as well as within a dataset. Practically data of the same group of objects are usually stored in different datasets; in each data set, the data dimensions are not necessarily the same and unreal data may exist. Fuzzy clustering of a single data set would bring about less reliable results. And these data sets cannot be integrated for some reasons. An interesting application of vertical clustering occurs when dealing with huge data sets. Instead of clustering them in a single pass, we split them into individual data sets, cluster each of them separately, and reconcile the results through the collaborative exchange of prototypes. Vertical collaborative fuzzy C-Means is a useful tool for dealing collaborative clustering problems where a feature space is described in different pattern-sets. In this paper we use collaborative fuzzy clustering, first we cluster each data set individually and then optimize in accordance with the dependency of these datasets is adopted so as to improve the quality of fuzzy clustering of a single data set with the help of other data sets, taking personal privacy and security of data into consideration. © 2013 Springer-Verlag Berlin Heidelberg.
Please use this identifier to cite or link to this item: