PCA and DWT Based Gene Selection Technique for Classification of Microarray Data

Classification of cancer is a vital research area in the bioinformatics area. The microarray technology allows the instantaneous observing of thousands of genes expressions in a single experiment that motivates the development of cancer classification. The major issue in microarray data is the high dimensionality problem, which happens due to great amount of genes existing when associated with fewer amounts of samples. In this paper, combination of dimensionality reduction technique and Wavelets are used to extract the informative genes from the dataset. The Principal Component Analysis (PCA) technique will reduce the dimension of the dataset, and then followed by the Discrete Wavelet Transform (DWT) will extract the most informative genes and those genes are forwarded to the classifiers namely k Nearest Neighbor (kNN) and Support Vector Machine (SVM) classifiers for classification. The proposed method is applied on the two widely accessible microarray datasets. Investigational outcomes show that the PCA and DWT constructed gene selection technique achieves high classification accuracy when compared with the conventional methods.