In the paper the clustering algorithms based on fuzzy set theory are considered. Modifications of the Fuzzy C-Means and the Possibilistic C-Means algorithms are presented, which adjust them to deal with data streams. Since data stream is of infinite size, it has to be partitioned into chunks. Simulations show that this partitioning procedure does not affect the quality of clustering results significantly. Moreover, properly chosen weights can be assigned to each data element. This modification allows the presented algorithms to handle concept drift during simulations. © 2012 Springer-Verlag Berlin Heidelberg.
CITATION STYLE
Jaworski, M., Duda, P., & Pietruczuk, L. (2012). On fuzzy clustering of data streams with concept drift. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 7268 LNAI, pp. 82–91). Springer Verlag. https://doi.org/10.1007/978-3-642-29350-4_10
Mendeley helps you to discover research relevant for your work.