Skip to main content

Advertisement

Fig. 4 | SpringerPlus

Fig. 4

From: Non-invasive detection of fasting blood glucose level via electrochemical measurement of saliva

Fig. 4

a The ROC plots for linear logistic regression model. The coordinates for the normal and diabetic populations were (0.69, 0.16) and (0.82, 0.31), respectively. b The ROC plots for the ANN model illustrating the coordinates to be closely placed at (0.84, 0.2) for the normal class and at (0.8, 0.16) for the diabetic population. c The ROC plots for linear-SVM. d The ROC plots for RBF-SVM models. The coordinates for the curves were (0.72, 0.16) for the normal class and (0.84, 0.28) for the diabetic class in linear, and (0.8, 0.1) for the normal class and (0.9, 0.2) for the diabetic class in RBF-SVM. The RBF-SVM ROC coordinates being closer to (1, 0) suggested a better fit than the linear model

Back to article page