Skip to main content
. 2019 Apr 7;2019:3491852. doi: 10.1155/2019/3491852

Figure 4.

Figure 4

The results of the important feature identification in the serum data. (a) The multivariate ROC curves based on the cross-validation to determine the features (5 features); (b) the predicted class probability plots (average of the cross-validation) for each sample using the base classifier (based on AUC with 5 features). (c) Rank features by the selected frequency of being selected.