Classification of Colon Cancer by using Support Vector Machines

V T Ram Pavan Kumar, P. L. Ramesh, M Arulselvi

Abstract: Colon cancer is considered a dangerous disease in humans, and it is one of the main risks to human life. In spite of the advances in screening, analysis, and handling, colorectal cancer (CRC) or colon cancer is the major widespread and Third-leading cause globally. The precise prediction of cancer with the gene data is very important for diagnosing cancer. However, the enormous dimensions of the gene expression data make the cancer prediction approach more complex. This paper devises a novel Support Vector Machine (SVM) for the classification of colon cancer. Here, the input data are gathered from the dataset and is fed to the feature selection module for selecting the features. Here, the selection is made using the Entropy and the Bhattacharya distance measures separately in order to select the unique features. Once the features are selected developed SVM provide the final classified output. The proposed SVM classifier outperformed other techniques with a maximum accuracy of 97.38%, higher sensitivity of 97.61%, and maximum specificity of 96.77% in terms of training data.

Keywords: Colon cancer, Gene expression data, Entropy measure, Bhattacharya distance, Support Vector Machine