| |
| In order to perform successful diagnosis and treatment of cancer,discovering and classifying cancer types correctly is essential. Most of the existing worksadopt single clustering algorithms to perform class discovery frombio-molecular data. Unfortunately, single clustering algorithms havelimitations, which are lack of the robustness, stableness andaccuracy. In this paper, we develop a new probabilistic subspaceensemble framework known as PSEFminer for cancer microarray dataanalysis. PSEFminer integrates the probabilistic subspace generator,the self-organizing map(SOM) and the normalized cut algorithm intothe ensemble framework to discover the underlying structure fromcancer microarray data. The experiments in cancer datasets show that($i$) the probabilistic subspace generator plays an important roleto improve the performance of PSEFminer; ($ii$) PSEFmineroutperforms most of the state-of-the-art cluster ensemble algorithmswhen applied to cancer gene expression data. |
|
| Keywords:Cluster ensemble; Class discovery; Cancer data |
|
