Table 2 The statistical p-value after testing the null hypothesis between different learning flows: IV, IC, SEV, SEC, and SVEVC considering three classes of TMA tissues and uncertain predictions in evaluation B.

From: Classification of colorectal tissue images from high throughput tissue microarrays by ensemble deep learning methods

 

Tumor

Normal

Other

Uncertain cases

IV versus IC

0.0156

0.2932

0.625

0.0235

IV versus SEV

4.15e−05

0.3711

0.0036

8.75e−05

IC versus SEC

8.35e−06

1

0.0020

3.82e−05

SEV versus SEC

0.0031

0.1422

0.4636

0.0227

IV versus SVEVC

4.52e−05

0.2807

0.0002

0.0048

IC versus SVEVC

1.89e−05

0.0421

0.0001

0.0133

SEV versus SVEVC

0.0007

0.5896

0.0033

0.0707

SEC versus SVEVC

0.1103

0.0417

0.0030

0.3448

  1. Using p values < 0.00625 as the statistically significant level of difference, we highlight the improvement of prediction cases in bold.