31P MRS Data Diagnosis of Hepatocellular Carcinoma Based on Support Vector Machine

SVM (support vector machine) is a new machine-learning technique which is developed based on statistical theory and it is applied in the various fields in recent years. We use SVM model based on ³¹P MRS (³¹Phosphorus magnetic resonance spectroscopy) data to distinguish three categories of hepatocellular carcinoma, hepatic cirrhosis and normal hepatic tissue. The recognition accuracy of the three categories was obtained, and the classification accuracy of SVM based on polynomial and radial basis function kernel is compared. The result of experiments shows that SVM model based on ³¹P MRS data provides diagnostic prediction of liver in vivo, and the performance based on polynomial is better than based on radial basis function kernel.