Table 3 Impact of the positive weight (W) in the loss function on ROC AUC, accuracy (Acc), PPA, PPV, F1 score (F1), and NPA across test and holdout sets

From: Predicting ROS1 and ALK fusions in NSCLC from H&E slides with a two-step vision transformer approach

 

Test

Holdout

W

AUC

Acc

PPA

PPV

F1

NPA

AUC

Acc

PPA

PPV

F1

NPA

5

0.85

0.92

0.50

0.06

0.10

0.92

0.85

0.92

0.48

0.06

0.10

0.93

10

0.85

0.88

0.58

0.05

0.09

0.89

0.85

0.89

0.58

0.05

0.09

0.89

20

0.86

0.84

0.72

0.04

0.08

0.84

0.85

0.84

0.68

0.04

0.07

0.84

30

0.85

0.79

0.79

0.04

0.07

0.79

0.85

0.79

0.72

0.03

0.06

0.79

  1. These values represent the average across 5 folds for the ROS1 fusion models.
Back to article page