Evolutionary Computational Algorithm by Blending of PPCA and EP-Enhanced Supervised Classifier for Microarray Gene Expression Data

Abstract

In DNA microarray technology, gene classification is considered to be difficult because the attributes of the data, are characterized by high dimensionality and small sample size. Classification of tissue samples in such high dimensional problems is a complicated task. Furthermore, there is a high redundancy in microarray data and several genes comprise inappropriate information for accurate classification of diseases or phenotypes. Consequently, an efficient classification technique is necessary to retrieve the gene information from the microarray experimental data. In this paper, a classification technique is proposed that classifies the microarray gene expression data well. In the proposed technique, the dimensionality of the gene expression dataset is reduced by Probabilistic PCA. Then, an Artificial Neural Network (ANN) is selected as the supervised classifier and it is enhanced using Evolutionary programming (EP) technique. The enhancement of the classifier is accomplished by optimizing the dimension of the ANN. The enhanced classifier is trained using the Back Propagation (BP) algorithm and so the BP error gets minimized. The well-trained ANN has the capacity of classifying the gene expression data to the associated classes. The proposed technique is evaluated by classification performance over the cancer classes, Acute myeloid leukemia (AML) and Acute Lymphoblastic Leukemia (ALL). The classification performance of the enhanced ANN classifier is compared over the existing ANN classifier and SVM classifier.