A Comparison of SVM-based Evolutionary Methods for Multicategory Cancer Diagnosis using Microarray Gene Expression Data

Selection of relevant genes that will give higher accuracy for sample classification (for example, to distinguish cancerous from normal tissues) is a common task in most microarray data studies. An evolutionary method based on generalization error bound theory of support vector machine (SVM) can select a subset of potentially informative genes for SVM classifier very efficiently. The bound theories are developed for binary SVM, however multiclass SVMs do not have established bounds on the generalization error. Several multiclass SVMs have been proposed where multiclass SVMs are typically constructed by combining several binary SVMs. We evaluate an estimate of a generalization error bound for a multiclass SVM by combining the error bound of binary SVMs which are used to construct the multiclass SVM. In this paper our aims are to compare the performance of several multiclass SVMs in the SVM-based evolutionary method and then find the best multiclass SVM classifier in the SVM-based evolutionary method for multicategory cancer diagnosis using microarray gene expression data.