Table 3 A summary of weighted k values and ICC Weighted kappa values less than 0.20 indicate poor agreement, 0.21 to 0.40 indicate fair agreement, 0.41 to 0.60 indicate moderate agreement, 0.61 to 0.80 indicate good agreement, and greater than 0.80 indicate very good agreement. ICC below 0.5 is considered poor reliability, 0.5 and 0.75 as moderate reliability, 0.75–0.9 as good reliability, and over 0.90 as excellent reliability31,32.

From: An AI based classifier model for lateral pillar classification of Legg–Calve–Perthes

 

DR2

DR3

DR4

CNN

DR5

DR6

Avg. ICC

95% CI

ICC rating

DR1

0.825

0.556

0.572

0.607

0.581

0.555

0.932

0.894–0.956

Excellent

DR2

 

0.547

0.642

0.605

0.626

0.598

0.929

0.890–0.954

Excellent

DR3

  

0.569

0.513

0.561

0.610

0.831

0.725–0.894

Good

DR4

   

0.478

0.732

0.783

0.859

0.656–0.930

Good

CNN

    

0.433

0.436

0.868

0.796–0.915

Good

DR5

     

0.774

0.853

0.744–0.912

Good

DR6

      

0.863

0.676–0.931

Good

  1. Significant values are in bold.