Classification of malignant mesothelioma cancer using support vector machine

Abstract

Researchers have prioritized to identify the disease in its premature stage, to control the invasive nature of cancer. The most prominent causes of cancer are environmental issues, life style and genetic heritage. Malignant Mesothelioma (MM) is one of the fastest growing neoplasm tumour in human body, that originates due to mesothelium cells in various parts of the human body, and directly affects the pleura. The main causes of MM are asbestos exposure, exposure to the high doses of radiation to the chest or abdomen, genetics disposition and the infection of simian virus 40. In this paper MM tumour classification is performed using Support Vector Machine (SVM). Tumour is classified as either malignant or benign. SVM is trained on features extracted in the form of symptoms of MM cancer. The proposed method is compared with Probabilistic Neural Network (PNN) classification method and Multi-layered neural networks (MLNN) and shows better results than both PNN and MLNN.