Logo Logo
Switch Language to German

Mautz, Dominik; Böhm, Christian and Plant, Claudia (2016): Subspace Clustering Ensembles Through Tensor Decomposition. IEEE 16th International Conference on Data Mining Workshops (ICDMW), Barcelona, Catalonia, Spain, 12-15 Dec. 2016. IEEE Computer Society. pp. 1225-1234

Full text not available from 'Open Access LMU'.


In recent years many different subspace clustering algorithms and related methods have been proposed. They promise to not only find hidden structures in data sets, but also to select for each structure the features, which are most prominent. Yet, most of these methods suffer from the same problem: finding a satisfactory clustering result heavily depends on an adequate configuration of the parameters. In case of insufficient parameterization, a result is potentially hard to interpret and might contain hundreds of clusters. For traditional clustering algorithms different ensemble methods have been developed, which mitigate these effects by incorporating multiple clustering outputs into a consensus result. However, most of these methods cannot be straightforwardly adopted to include subspace information. We propose a novel subspace clustering ensemble algorithm SubCluEns(1) based on the minimum description length principle. It allows combining multiple results of subspace and projected clustering algorithms into a consensus clustering.

Actions (login required)

View Item View Item