Prasun Chakrabarti , Manish Tiwari , Tulika Chakrabarti Abstract—The paper deals with performance vector analysis based upon efficient classification techniques for liver cancer investigation. The classification algorithms applied herewith include Naïve Bayes classifier and Support Vector Machines with different features such as SGOT, SGPT, and Alkaline Phosphates. The role of Chemistry has also been pointed out in liver cancer diagnosis. The entire work has a significant impact on medical field and society as well. Index Terms—Naïve Bayes classifier, Support Vector Machine, SGOT, SGPT, Alkaline Phosphates I. INTRODUCTION In 20 th century ,cancer is one of the uncured diseases in human body. Cancer is describing as uncontrolled growth of cells that attack on healthy tissues [1]. Researchers have given many proposals on cancer classification. One of the cancer diseases is Hepatocellular carcinoma (HCC), Predication of HCC is difficult as information of liver cancer is known at a later stage [2-3]. Hepatcellular carcinoma is most common type of liver cancer that affects male more than females [4]. According to National Health Service (NHS) [5], 1,500 people die from HCC in United Kingdom. According to World Health Organization (WHO) [6], Death rate in the parts of Africa and Eastern Asia is particularly high. HCC is cause of regular high alcohol consumption, having unprotected sex and injecting drugs with shared needles [7-8]. Symptoms of Hepatocellular carcinoma (HCC) are jaundice, abdominal pain, unexpected weight loss etc. Classification techniques are popular in various automatic medical diagnosis tools. Liver disease can be diagnosed by analyzing level of enzymes in the blood [9]. ___________________________________________________________ Prof. Prasun Chakrabarti belongs to the Department of Computer Science and Engineering, Sir Padampat Singhania University, Udaipur, Rajasthan, India (e- mail: prasun.chakrabarti@spsu.ac.in). Manish Tiwari belongs to the Department of Information Technology, Geetanjali Institute of Technical Studies, Udaipur, Rajasthan, India (e-mail: immanishtiwari@gmail.com). Dr. Tulika Chakrabarti belongs to the Department of Chemistry, Sir Padampat Singhania University, Udaipur, Rajasthan, India (e-mail: tulika.chakrabarti@spsu.ac.in). II. METHODOLOGY Classification algorithms such as Naïve Bayes classifier, K- nearest neighbor and Support Vector Machines etc. have been used. The datasets used for liver cancer analysis are mainly BUPA Liver Disorders datasets taken from University of California at Irvine (UCI) Machine Learning Repository[10]. Classification is a two phase process - one for training set of tuples and second phase model for classification and its performance is analyzed with the testing set of tuples[11]. Bayesian classifier works on one assumption viz. the effect of an attribute value on a given class is independent of the values of the other attributes. This assumption is called class conditional independence. A support vector machine(SVM) algorithm[11] is the machine learning classifier which classifies data into two categories of performing classification and constructing N-dimensional hyper plane. It is based on Vapnik- theory(VC) and structural risk minimization(SRM) that has the aim to find out the trade-off between minimizing the training set error and maximizing the margin, in order to achieve the best generalization ability and remains resistant to over fitting. For cancer classification, classes have been demarked as benign and malignant. A transformed attribute used to define the hyper plane is called a feature. Rapid Miner Ver. 5 has been used for performance analysis and simulation purpose. III. RESULTS Dataset BUPA liver disorder dataset ( University of California at Irvine (UCI) machine learning repository) chosen for implementation has 6 attributes and 345 instances. First five attributes belong to blood tests that are sensitive to liver disorder resulting from excessive alcohol consumption. Implementation of the SVM(evolutionary) model is based on genetic approach and it is the part of data transformation . After that the cross validation has been applied, 10 folds in the Performance Vector analysis in context to liver cancer – A Support Vector Machine Approach with a survey on the latest Perspectives of Chemistry in liver cancer treatment International Journal of Computer Science and Information Security (IJCSIS), Vol. 14, No. 9, September 2016 1238 https://sites.google.com/site/ijcsis/ ISSN 1947-5500