What is $y$? Why is $k$ the ceil of $n/2$? What is $y \geq k$?
$\begingroup$
$\endgroup$
2
-
$\begingroup$ Where did you get this text/screenshot from? Which book, article, etc.? $\endgroup$– nbroDec 4, 2020 at 11:19
-
$\begingroup$ This SE:AI post will shed light into the application of the binomial probability to this problem of ensembles. $\endgroup$– Snehal PatelDec 25, 2022 at 4:01
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
Y ensemble size voting wrong
k = 50% or majority threshold
If you have 11 models. Then the majority of models is anything bigger than 50% of the number of ensemble models. In the example where you have 11 base models. The majority would be anything bigger than N/2 or 11/2. But since 11 is an odd number and cannot be divided by 2. We have to use the python ceiling function to round 5.5 to 6. The in other words. For your ensemble to be wrong. We must look for the probability that Y of them are wrong. Y is ≥ than K and K is ceil(N/2). Thus we must calculate this probability(6 wrong) like shown above summating the possibility of each of these discrete combinations into one