Support Vector Machine Prediction Model Based on ROC Technology and Application

It is deficiency to use accuracy as a measurement to evaluate model classifying ability. This paper proposes a measurement method which uses the area under the ROC curve, or AUC value, to evaluate the performance of the model. Furthermore, applying cross validation and grid-search methods, through designed algorithms, to build an optimization of support vector machines medical prediction model. The model was applied to diagnosis of predicting of coronary heart disease. The results show that the model has the characteristics of global optimization and easy to implement.