Eegore wrote on 07/24/20 at 21:43:19:
"...The test predicted with 100% accuracy all of the negative samples, and 29 of 30 positive samples were predicted accurately ..."
Couple of questions;
How was the, ‘test’, ‘tested’ ?
Why do you believe that 3.3 out of 100 wrong ‘tests’ are good ?
Which would be, (how many ?), wrong out of 4 MILLION cases reported, (case being a positive, regardless of anything else)
Why is the word,
‘Predicted’, used to gauge the test ?
In my experience,
‘predicted’, is what, ‘could/should’, be.
Like a certain amount of powder, of a certain kind, in a certain case, with a certain bullet weight, which has a certain diameter, which has a certain coating, which is made of a certain material, which has a certain length. It is/can be,
‘predicted’, to propel that projectile at ‘X’ speed.
A, ‘test’, is absolute. In the above info, the actual shooting, then testing, will determine the actual FPS. Then in the real world, one can load something by,
‘prediction’, that it will above or below a certain FPS/FPE. Yet only a, ‘test’, can prove it is, or not.
Why do you believe a, ‘prediction’, is a ‘test’ ?