Skip to main content

Table 2 SE-ResNet performance for three different experimental settings

From: Pattern recognition of topologically associating domains using deep learning

Methods

TPR

TNR

Precision

Acc.

F1

AUC

Same species, cross cell types

      

 SE-ResNet_Human.ES

0.922

0.832

0.846

0.877

0.882

0.877

 SE-ResNet_Mouse.ES

0.853

0.763

0.783

0.808

0.817

0.808

Cross cell types

      

 SE-ResNet_ES

0.907

0.841

0.851

0.874

0.878

0.874

Cross species

      

 SE-ResNet_Human

0.876

0.864

0.862

0.870

0.869

0.870

 SE-ResNet_Mouse

0.837

0.900

0.893

0.868

0.864

0.868