How Do I Use A Trained Ensemble Model In Optimization Toolbox?

I have created a classification ensemble based on gentle boosting. I'm trying to figure out how to use this with the optimization toolbox, with the idea being to use the solver to optimize the predicted score for a given class.

Presumably, I have to use the underlying structure of the ensemble model to create the objective function, with the objective being to optimize the score. Some of the variables will have constraints and others will be fixed.

First, is this reasonable, and second, how do I extract the underlying model from the trained ensemble to use in my objective function?

Thanks.

Answer 1

[Alan Weiss]

I don't understand what you mean. Boosting is a method of optimization that tries to minimize the "loss". I mean, creating a boosting ensemble is a method of optimization, and I do not understand why you would want to run a second round of optimization. For a brief description of the GentleBoost algorithm, see GentleBoost.

Alan Weiss

MATLAB mathematical toolbox documentation