Ensemble Classifiers based on Kernel ICA for Cancer Data Classification

Now the classification of different tumor types is of great importance in cancer diagnosis and drug discovery. It is more desirable to create an optimal ensemble for data analysis that deals with few samples and large features. In this paper, a new ensemble method for cancer data classification is proposed. The gene expression data is firstly preprocessed for normalization. Kernel Independent Component Analysis (KICA) is then applied to extract features. Secondly, an intelligent approach is brought forward, which uses Support Vector Machine (SVM) as the base classifier and applied with Binary Particle Swarm Optimization (BPSO) for constructing ensemble classifiers. The leukemia and colon datasets are used for conducting all the experiments. Results show that the proposed method produces a good recognition rate comparing with some other advanced artificial techniques.