Skip to main content

Table 3 Performances of SVM classifiers

From: A study of hepatic fibrosis staging methods using diffraction enhanced imaging

Kernel function Kernel scale Accuracy Sensitivity specificity Training time (s) Prediction speed (obs/s)
Linear 0.07 93.83% 94.32% 91.90% 2367.9 170,000
Quadratic 0.37 99.31% 99.66% 97.93% 1655.7 72,000
Cubic 0.78 99.57% 99.69% 99.10% 699.22 76,000
Gaussian 0.55 99.49% 99.73% 98.14% 93.538 6900