Paper Title
Comparative Analysis of Support Vector Machine using Various Kernel Methods – An Experimental Approach

Abstract
SVM, a supervised data mining technique, has retained the efficacious and conspicuous place to predict diverse diseases, or to predict the results of students or to forecast the weather conditions or any prediction based on the respective datasets, with acceptable and satisfactory performance results. Each data item is represented as a point in n-dimensional space of SVM.SVM model defines and divides the testing datasets very competently into the respective classes using linear or non-linear methodology. In this research work an empirical and realistic study is performed to compare and analyze the performance results of support vector machine using three different kernel methods. RBF kernel, polynomial kernel and Pearson VII kernel (PUK) methods are the three engaged with SVM in disjoint fashion. Experimental work is performed in WEKA simulation tool using online available the Hungarian datasets. SVM with RBF kernel overpowers the performance results shown by other two kernel methods when assimilated with SVM.