Skip to main content
. 2020 Jul 22;10:1186. doi: 10.3389/fonc.2020.01186

Table 2.

The performance comparison of different models.

No. of classes Model Validation loss Test accuracy
Start End
2 Lung-DL model 0.61 0.29 89.52%
Dense model 0.79 0.3 81.85%
3 Lung-DL model 0.91 0.34 92%
Dense model 1.15 0.36 72.91%

The 2-Class classification taskis used to distinguish the invasiveness degree. The 3-Class classification task is used to predict the micropapillary component.