Prediction of the expected improvement after getting the Bayesopt result output.

I conducted a Bayesian optimization process using the Bayesopt function, which concluded after 250 iterations. The process generated an output file containing various pieces of information. From this file, I understand that predictions at any given point can be made using the "predictObjective" or "predictConstraints" functions. However, I am interested in predicting the current values of the acquisition function used during the Bayesopt process. Is there a method to achieve this?

Answer 1

[Umang Pandey]

Hi,

From what I understand, you want to obtain the acquisition function values while implementing "bayesopt".

Refer to the following documentation which explains the "Bayesian Optimization Algorithm" that might be of help to you :

https://www.mathworks.com/help/stats/bayesian-optimization-algorithm.html

Hope this helps!

Best,

Umang