Prominent feature selection of microarray data

For wavelet transform, a set of orthogonal wavelet basis aims to detect the localized changing features contained in microarray data. In this research, we investigate the performance of the selected wavelet features based on wavelet detail coefficients at the second level and the third level. The genetic algorithm is performed to optimize wavelet detail coefficients to select the best discriminant features. Experiments are carried out on four microarray datasets to evaluate the performance of classification. Experimental results prove that wavelet features optimized from detail coefficients efficiently characterize the differences between normal tissues and cancer tissues.