Fig. 4.

The performance evaluation of the devised network on the wild-type CLASH dataset. The solid green/orange/blue line represents the mean results on the training/validation/test sets in the 30 split runs. And the colored dashed line depicts the variation of the results in the 30 split runs. a The randomness of the 30 ten-fold splits was ensured by the average 24/3/3 selected times of each pair to be used in the training/validation/test sets. b Learning curves for our devised model. c The ROC curve comparison of the validation results and the test results of the devised model