Discover MakerZone

MATLAB and Simulink resources for Arduino, LEGO, and Raspberry Pi

Learn more

Discover what MATLAB® can do for your career.

Opportunities for recent engineering grads.

Apply Today

Thread Subject:
What Regression and Correlation Models to use for Kriging

Subject: What Regression and Correlation Models to use for Kriging

From: Thomas Clark

Date: 29 Oct, 2012 15:07:09

Message: 1 of 2

Hi all,

I'm using the DACE toolbox...
http://www2.imm.dtu.dk/~hbni/dace/dace.pdf
... to perform Kriging on a design space with 1000 points in 12 dimensions.

The toolbox has built in regression and correlation models, but I need to select which to use. Moreover, I need to make initial estimates for parameters of the models (theta on page 21 of this document).

After hours of trawling the net, I simply cannot find any guide for which correlation and regression model I should choose.

Should I simply try every combination to find the one with least MSE? Will that then be the best model to use, or are there other considerations? Or should my selection be based on physical intuition about the parameters, as in a curve fit?
How sensitive is the fit likely to be to starting point estimation of the theta parameter?

Any guidance truly appreciated!

Best,

Tom

Subject: What Regression and Correlation Models to use for Kriging

From: mouna

Date: 17 Aug, 2013 11:23:07

Message: 2 of 2

Hi to all,

I have the same problem , I want to use DACE toolbox and I don't know which Regression and Correlation Models should I use?

Many thanks,

"Thomas Clark" wrote in message <k6m62t$slh$1@newscl01ah.mathworks.com>...
> Hi all,
>
> I'm using the DACE toolbox...
> http://www2.imm.dtu.dk/~hbni/dace/dace.pdf
> ... to perform Kriging on a design space with 1000 points in 12 dimensions.
>
> The toolbox has built in regression and correlation models, but I need to select which to use. Moreover, I need to make initial estimates for parameters of the models (theta on page 21 of this document).
>
> After hours of trawling the net, I simply cannot find any guide for which correlation and regression model I should choose.
>
> Should I simply try every combination to find the one with least MSE? Will that then be the best model to use, or are there other considerations? Or should my selection be based on physical intuition about the parameters, as in a curve fit?
> How sensitive is the fit likely to be to starting point estimation of the theta parameter?
>
> Any guidance truly appreciated!
>
> Best,
>
> Tom

Tags for this Thread

What are tags?

A tag is like a keyword or category label associated with each thread. Tags make it easier for you to find threads of interest.

Anyone can tag a thread. Tags are public and visible to everyone.

Contact us