Microarray Colon Data Classification using Hybrid Classifiers

The major cause of cancer disease is due to genetic disorders and the Microarray technology has been developed to analyze the expressions for more than thousands of gene samples at a time. The Microarray data is of high dimension and it contains noise that affects the classification accuracy. This paper has applied eight different classifiers to a human colon cancer dataset to analyze the performance of classification. The eight different classifiers namely K-Nearest Neighbours (KNN), Support Vector Machine (SVM), Linear Discriminant Analysis (LDA), Log Regression based LDA, Exponential LDA, Hybrid LDA, Adaboost, Artificial Bee Colony- Particle Swarm Optimization (ABC-PSO) classifiers are used. In this paper, four types of hybrid classifiers are proposed. Then the existing classifier and hybrid classifiers performance are analyzed. The obtained results illustrate the accuracy of 100% is obtained with Log regression LDA and Hybrid LDA classifiers. The classifier’s advantages are that it produces good classification accuracy and misclassification errors reduction when compared with the other conventional classifiers.