On Constructing Clusters from Non-Euclidean Dissimilarity Matrix by Using Rough Clustering
In this paper, we present a clustering method which can construct interpretable clusters from a dissimilarity matrix containing relatively or subjectively defined dissimilarities. Experimental results on the synthetic, numerical datasets demonstrated that this method could produce good clusters even when the proximity of the objects did satisfy the triangular inequality. Results on chronic hepatitis dataset also demonstrated that this method could absorb local disturbance in the proximity matrix and produce interpretable clusters containing time series that have similar patterns.
KeywordsEquivalence Relation Cluster Method Cluster Result Good Cluster Dissimilarity Matrix
Unable to display preview. Download preview PDF.
- 1.Berkhin, P.: Survey of clustering data mining techniques. Accrue Software Research Paper (2002), http://www.accrue.com/products/researchpapers.html
- 2.Everitt, B.S., Landau, S., Leese, M.: Cluster Analysis, 4th edn. Arnold Publishers (2001)Google Scholar
- 3.Hirano, S., Tsumoto, S.: An indiscernibility-based clustering method with iterative refinement of equivalence relations. Journal of Advanced Computational Intelligence and Intelligent Informatics 7, 169–177 (2003)Google Scholar