Stroke Risk Prediction through Non-linear Support Vector Classification Models

The aim of this study is to find the possible risk of Cerebro Vascular Accident (CVA) or Stroke by subjecting the risk factors toSupport Vector Machines (SVM).The prediction of the attack of the disease is highly dependent on the quantification of risks contributed byeach factor. Therefore an assessment of relative intensity of risk contributed by the factors is imperative for early prediction and preventablemeasures. The classification accuracies are achieved through the efficient kernel functions of Radial Basis Function (RBF=98%) and Polynomial (Poly=92%) and finally these results are compared with benchmarking evaluation methods like classification accuracy, sensitivity, specificityand confusion matrix. The proposed stroke risk prediction models are obtained with satisfactory accuracy and it would be promising models inthe classification of stroke risk prediction process