Dual edge classifier based support vector machine (Desvm) classifier for clinical dataset

Abstract

Data mining is the progression of determining hidden information that are available in the existing data. Data mining discovers interesting, convenient relationships in huge volume of data. Many fields including medical field is using data mining for classifying the data. Classification is method which assigns a data in the collection to predict the objective class. Classifying a diabetic patient is tedious job in the current medical field. The main intention of this paper is to propose a novel classifier enhancing support vector machine to correctly classify the diabetic patients more accurately than the previous classifiers. Performance metrics such as sensitivity, specificity, rate of true positive and false positive, precision, accuracy and time taken for feature selection are used. In the proposed classifier threshold value is fixed for metric recall and true negative rate. The results are demonstrated with better performance.