Abstract | ||
---|---|---|
Constrained clustering has received substantial attention recently. This framework proposes to support the clustering process by prior knowledge in terms of constraints (on data items, cluster size, etc.). In this work we introduce clustering combination into the constrained clustering framework. It is argued that even if all clusterings of an ensemble satisfy the constraints, there is still a need of carefully considering the constraints in the combination method in order to avoid a violation in the final combined clustering. We propose an evidence accumulation approach for this purpose, which is quantitatively compared with constrained algorithms and unconstrained combination methods. |
Year | DOI | Venue |
---|---|---|
2009 | 10.1007/978-3-642-03070-3_27 | MLDM |
Keywords | Field | DocType |
evidence accumulation approach,clustering process,constrained clustering combination,data item,unconstrained combination method,final combined clustering,clustering combination,combination method,clustering framework,prior knowledge,cluster size,satisfiability | Data mining,Fuzzy clustering,CURE data clustering algorithm,Computer science,Consensus clustering,Artificial intelligence,Cluster analysis,Canopy clustering algorithm,Correlation clustering,Pattern recognition,Constrained clustering,Brown clustering,Machine learning | Conference |
Volume | ISSN | Citations |
5632 | 0302-9743 | 1 |
PageRank | References | Authors |
0.36 | 13 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Daniel Duarte Abdala | 1 | 16 | 2.73 |
Xiaoyi Jiang | 2 | 2184 | 206.38 |