Timeline for Optimizing a "black box" function: Linear Regression or Bayesian Optimization... what's the difference?
Current License: CC BY-SA 3.0
9 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Sep 28, 2016 at 16:09 | comment | added | Sycorax♦ | @GeoMatt22 Thank you for that additional background! | |
| Sep 28, 2016 at 15:53 | comment | added | GeoMatt22 | The OP is essentially talking about building a surrogate model. As implied by this answer, the polynomial approach is typically used for local surrogates, while Gaussian processes are used for global surrogates (e.g. DACE). It is worth noting that for local models, quadratic trust region methods are typical (not higher order!). The reference I give here has some examples. | |
| Sep 28, 2016 at 15:39 | vote | accept | TravisJ | ||
| S Sep 28, 2016 at 15:39 | history | suggested | TravisJ | CC BY-SA 3.0 |
Added a link to the referenced paper.
|
| Sep 28, 2016 at 15:39 | review | Suggested edits | |||
| S Sep 28, 2016 at 15:39 | |||||
| Sep 28, 2016 at 15:36 | history | edited | Sycorax♦ | CC BY-SA 3.0 |
added 40 characters in body
|
| Sep 28, 2016 at 15:35 | comment | added | Sycorax♦ | @TravisJ Yes, that's the one. I just found the page I was looking for p. 464. | |
| Sep 28, 2016 at 15:31 | comment | added | TravisJ | If the Jones paper is the "Efficient Global Optimization of Expensive Black-Box Functions" then the link you provided in the comments links to that paper. | |
| Sep 28, 2016 at 15:26 | history | answered | Sycorax♦ | CC BY-SA 3.0 |