Table 6 Ablation study of the proposed model.

From: A deep learning model for epidermal growth factor receptor prediction using ensemble residual convolutional neural network

Variant

Acc (%)

Sn (%)

Sp (%)

PR-AUC (%)

MCC

Plain CNN

90.80

91.90

89.60

95.10

0.771

RCNN (Residual CNN without ensemble)

91.50

92.40

90.20

95.80

0.784

RCNN with RBB-1 only

92.20

93.50

90.80

96.40

0.801

RCNN with RBB-2 only

92.00

92.80

91.10

96.20

0.797

Full ERCNN (Proposed)

93.48

94.53

92.58

97.38

0.816

Back to article page