Sensitivity/specificity on the training and testing sets for different classification models with the optimal parameter combination.

<p>“▴” denotes that a p-value of <0.001 was obtained by ANOVA between F-score_FDA and F-score_SVM; “<b>*</b>” denotes that a p-value of <0.001 was obtained by ANOVA between F-score_BPNN and F-score_SVM; for BPNN, the number of hidden nodes = 5, and the learning rate = 0.03; for SVM, radial = 32, and penalty parameter <i>C</i> = 2⁸.</p><p>Sensitivity/specificity on the training and testing sets for different classification models with the optimal parameter combination.