Diagnostic Method of Diabetes Based on Support Vector Machine and Tongue Images

Table 2

SVM classification before and after data processing.

Datasets	Accuracy of cross-validation (%)	Accuracy of training samples (%)	Accuracy of test samples (%)	Running time (s)
Raw data	81.65	100.00	77.83	817.86
Normalized data	84.35	99.53	78.77	747.40
Normalized data with PCA	83.06	99.88	79.72	465.52