Neural network based classification of Glioma Grade III and Grade IV

Abstract

Microarray technology is a widely accepted for cancer subtype detection. Microarray gene expression data being a very high dimensional data, the greater challenge for microarray analysis is to identify the optimal set of genes for the purpose of classification. In this paper, we suggest a combination of ratio of mean values of a particular gene of both the classes, t- statistics and standard deviation to obtain the optimal subset of genes. The size of this data is further cut down with the help of wavelet transform. Finally, the classification is performed using neural network algorithms.