Adaptive model performance
Could you please clarify the answer to one of the questions in the "Predictive Analytics" Quiz in the Pega Decision Management Essentials 7.1 course?
The question is as follows:
What is the performance of an adaptive model if it CANNOT differentiate interested customers from not interested customers better than random selection? (Choose One)
0
0.5
50
1
The answer is 0.5 according to the quiz but I thought that the default performance is displayed as 50.00%.
Also the transcript contains the text: "And of course, we don’t yet know the Performance of the model, so that column shows the default, which is 50%. As soon as the model makes a bad prediction, the performance goes below 50%"
......So I thought the answer would be 50 but it is apparently 0.5. Have I mis-understood? If so, why is the answer 0.5?
Thanks in advance....
Pegasystems Inc.
NL
You are absolutely right, the correct answer is 50%. But 50% = 0.5 :-). And 50 != 0.5. That is why the answer is 0.5 and not 50.
The important bit to understand is that performance of a model is always a value between 0 and 1.
Hope this helps.
Miki