Cancer Classification from DNA Microarray Data using mRMR and Artificial Neural Network

Abstract

Cancer is the uncontrolled growth of abnormal cells in the body and is a major death cause now a days. Cancer may arise anywhere in the human body, and it names are remarked as body parts such as colon cancer, lung cancer, breast cancer. It is notable that cancer treatment is much easier in the initial stage rather than it outbreaks. DNA microarray based gene expression profiling has become efficient technique for cancer identification in early stage and a number of studies are available in this regard. Existing methods used different feature selection methods (e.g., wrapper and filter approaches) to select relevant genes and then employed distinct classifiers (e.g., artificial neural network, Naive Bayes, Decision Tree, Support Vector Machine) to identify cancer. This study considered information theoretic based minimum Redundancy Maximum Relevance (mRMR)method to select important genes and then employed artificial neural network (ANN) for cancer classification. Proposed mRMR-ANN method has been tested on a suite of benchmark data sets of various cancer. Experimental results revealed the proposed method as an effective method for cancer classification when performance compared with several related exiting methods.