A Two-Step Feature Selection Method to Predict Cancerlectins by Multiview Features and Synthetic Minority Oversampling Technique

Table 4

The prediction performance trained with the 13 optimal features and the prediction performance trained with the 13 features that are randomly selected from original features.

Method	Sensitivity	Specificity	Accuracy	MCC	AUC
Randomly selected features	0.631	0.609	0.620	0.240	0.659
Optimal features	0.779	0.717	0.748	0.497	0.787