Analysis of Classification Methods for Gene Expression Data

Abstract

The discovery of diseases at a molecular level is a great challenge for researchers in the field of bioinformatics and cancer classification. Understanding the genes that contribute to the cancer malady is a great challenge to many researchers. Cancer classification based on the molecular level investigation has gained the interest of researches as it provides a systematic, accurate and objective diagnosis for different cancer types. This Paper aims to present some classification methods for gene expression data. We compared the efficiency of three different classification methods; support vector machines, k-nearest neighbor and random forest. Two publicly available gene expression data sets were used in the classifications; Freije and Philips dataset. By performing the classification methods, results revealed that the best performance was achieved by using support vector machine classifier for both datasets comparing with other used classifiers.