Table 2 p-value results with FDR correction for the statistical comparison of classifier performance on the resampledRad hold-out test set.

From: Value of handcrafted and deep radiomic features towards training robust machine learning classifiers for prediction of prostate cancer disease aggressiveness

 

stableRad1

rad1

rad2

avgRad

unionRad

intersectionRad

rad1

0.0631

     

rad2

0.0466

0.7784

    

avgRad

0.6003

0.1511

0.1799

   

unionRad

0.7784

0.1624

0.1802

1.0000

  

intersectionRad

0.7784

0.0097

0.0223

0.1624

0.2958

 

resampledRad

0.0321

1.0000

1.0000

0.0969

0.0969

0.0097

  1. p-values< 0.05 are highlighted in bold. The column and the row corresponding to resampledRad and stableRad1, respectively, have been removed due to not containing information.
Back to article page