1
$\begingroup$

I created a boosting tree and got the probability for each tuple in my testing set. But I'm confused on how to combine each probability. Can someone tell me how to combine the probabilities?

Adaboost

$\endgroup$
  • $\begingroup$ You have to have the real $Y$ value $\endgroup$ – Juan Esteban de la Calle Apr 29 '19 at 2:50
  • $\begingroup$ I'm voting to close this question as off-topic because this question should be asked in cross validated $\endgroup$ – Mark.F Apr 29 '19 at 5:59
  • $\begingroup$ I don't understand. What do you mean? @Mark.F $\endgroup$ – rrr ppp Apr 29 '19 at 12:17
  • $\begingroup$ what do you mean by real value? the ground truth? I have. -1 or 1. @JuanEstebandelaCalle $\endgroup$ – rrr ppp Apr 29 '19 at 12:17
  • $\begingroup$ Yes, is against that variable what you compare the probability with (normally is 0,1) $\endgroup$ – Juan Esteban de la Calle Apr 29 '19 at 13:00
1
$\begingroup$

Is an interesting question.

You are solving a optimization problem

$max$ $AUC(\alpha_1*X_1 + \alpha_2*X_2,Y)$

$s.t.$ $\sum^2_{i=1}\alpha = 1$

You are maximizing the AUC using the models 1 and 2 (which are $X_1$ and $X_2$) subject to the sum of weights is 1.

You may any optimization algorithm, even Excel's solver to do this.

| improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.