Abstract
In this paper a new method for clustering is introduced that works fine for linearly inseparable clusters. This algorithm uses the of patterns’ concentration specification. We consider each pattern as a heat source to determine concentration by heat effect. Because of high concentration in clusters, heat in the clusters is higher than out of them, and this fact is used to determine and bound the clusters. CLIC algorithm can be implemented for any linearly inseparable cluster. The classification rate of this algorithm is higher than previously known methods (k-means, fuzzy c-means, etc). Finally, we cluster real iris data by using CLIC algorithm.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Osipova N (2004). Classification and clustering methods development and implementation for unstructured documents collections. Department of Programming Technology, Faculty of Applied Mathematics and Control Processes, St. Petersburg State University
Rose K (1998) Deterministic annealing for clustering, compression, classification, regression, and related optimization problems
Wallace M, Kollias S (2004) Robust, generalized, quick and efficient agglomerative clustering
Wolkenhauer Olaf Fuzzy classification, the iris-and-admission data sets
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2009 Springer Science+Business Media, LLC
About this paper
Cite this paper
Shahdoosti, H.R., Khayat, O. (2009). Clustering the Linearly Inseparable Clusters. In: Mastorakis, N., Mladenov, V., Kontargyri, V. (eds) Proceedings of the European Computing Conference. Lecture Notes in Electrical Engineering, vol 28. Springer, Boston, MA. https://doi.org/10.1007/978-0-387-85437-3_17
Download citation
DOI: https://doi.org/10.1007/978-0-387-85437-3_17
Published:
Publisher Name: Springer, Boston, MA
Print ISBN: 978-0-387-84818-1
Online ISBN: 978-0-387-85437-3
eBook Packages: EngineeringEngineering (R0)