nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

58325_en-US

liliconcom

Finding the correct factors

Test Diagram.PNG

Hi,

If i have a correlation as shown above whereby the F is a response for D & E but the factor D is actually a response of A & B. How do i construct a DOE experiment. This is what i wanted to ask, is it possible for the DOE table to generate this kind of relationship

Rgrds

Irfan

Re: Finding the correct factors

No, but you can see Help > Books > Profilers for lots of information and examples.

Re: Finding the correct factors

Hi Mark,

Do you have any example on how to use this in the graph profiler?

Rgrds

Irfan

Re: Finding the correct factors

Hi Mark,

Yes D, E and F are outcomes that we can measure

Irfan

Re: Finding the correct factors

Hi Phil,

Thanks for the answer, yes I will only get the F value when i perform experiment on D & E.

Thanks

Re: Finding the correct factors

I like Phil's answer but would you please clarify for me: are D, E, and F outcomes that you measure? If so, then follow's Phil's advice but model D and E as a function of A-C. Then model F as a function of the two models (D and E).

The Graph > Profiler will be particularly useful if this is the case. You can save the best model for D and for E. Then save the model for F. You can now profile F versus D and E or versus A, B, and C.

Re: Finding the correct factors

Really interesting.

I think you have to design an experiment for A, B and C. I assume you can't know the correlation of D and E without running the A, B, C experiment, so you don't know what settings of D and E will be possible.

One question: is the measurement of F a separate step in the experiment? Or would you get values of F when you measure D and E anyway?

An option is to create and run an experiment for A, B and C. Measure D and E and fit models. Then you can use D and E as covariate factors in a new Custom design. The Custom design would choose the available settings of D and E to maximise information. You can fit a model for F vs D, E. You can also nest the models for D and E in the model for F, so that you have the model of F as a function of A, B, C.

The model you described looks like Structural Equation Modelling (SEM). This is not possible with JMP currently. But you might want to look into that topic.

I hope that helps.

Phil

Re: Finding the correct factors

Anyone any advice on this?

Comments

There are no comments yet