Subscribe Bookmark RSS Feed

Individual Intervals of an exponential Fit Model

peter_michel

Community Trekker

Joined:

Apr 19, 2016

Hi,

can someone tell me how I get “95% Individual Intervals” of an exponential fit Model?

I need to know the upper Individual Interval at  certain x's. I have variance heterogeneity, hence I need the function/data vs the x-axis.

I’m using JMP13 - no Pro.

I prefer an interactive explanation

 

Thanxs Peter

1 ACCEPTED SOLUTION

Accepted Solutions
David_Burnham

Super User

Joined:

Jul 13, 2011

Solution

From the nonlinear platform you can select the following from the red triangle:

save formulas>save prediction formula

save formulas>save indiv confid limit formula

These formulae (formulas, really?) can then be plotted with the data in graph builder.

-Dave
8 REPLIES
David_Burnham

Super User

Joined:

Jul 13, 2011

Solution

From the nonlinear platform you can select the following from the red triangle:

save formulas>save prediction formula

save formulas>save indiv confid limit formula

These formulae (formulas, really?) can then be plotted with the data in graph builder.

-Dave
peter_michel

Community Trekker

Joined:

Apr 19, 2016

Hello David,

That is what I expected from JMP. But as you see in the screen dump there is no "save ind. Conf limits" point in my JMP

[cid:image001.png@01D322F8.B592FFE0]
Is this point missing in JMP and just available in Pro?
Is that a problem with my preferences?
Is the choice depending on the Model? Different Models have different choices available?
????
Many thanks
Peter

David_Burnham

Super User

Joined:

Jul 13, 2011

Peter

It looks like the image didn't come through so I can't see the screen shot.  Please try again, that will give us some clues!  I don't think it's a Pro issue, but version 13 comes with 2 flavours of nonlinear modelling: the original nonlinear platform and the new fit curve platform; the later doesn't have an option to save confidence intervals (although it you are a DIY fan you can save standard errors).

 

-Dave
peter_michel

Community Trekker

Joined:

Apr 19, 2016

David_Burnham

Super User

Joined:

Jul 13, 2011

FYI, my comments related to the Nonlinear platform, not the Fit Curve platform.

-Dave
peter_michel

Community Trekker

Joined:

Apr 19, 2016

Hello David,

Thanxs for taking your time with my question.

At least with my Version the print out of "Fit curve" and "Nonlinear" are very similar. The Fit Curve appears with the Model comparison closed. I attached both screen dumps in the PPTx. But there is no difference in the Save Formula Pull down. It seems very simple for you to get the prediction Intervalls and not the confidence intervals and difficult to me. That appears to me that I have a preference Problem that I don't see, that you have JMP Pro and I JMP or I have a Problem understanding the english in JMP. Is it really the wording you mentioned above "save formulas>save indiv confid limit formula"?

What else could cause the problem?

 

Thanxs Peter

susan_walsh1

Staff

Joined:

Jul 22, 2014

Peter,

In both of your screen shots you are actually using the Fit Curve platform. If you select Nonlinear, but assign a column without a formula to the X role, JMP will still use the Fit Curve platform. I would suggest that you choose Save Prediction Formula from the results you have. Then select the Nonlinear platform, but assign the newly save prediction formula column to the X role and the appropriate variable to the Y role. This will bring you to the Nonlinear platform that David was referring to, where you should be able to save the prediction intervals.

 

Sue

peter_michel

Community Trekker

Joined:

Apr 19, 2016

Thanks, chosing a formula as x, let the right button appear under save ....
Peter