Subscribe Bookmark RSS Feed

4PL fit (4 parameter logistic)

r_o_pedersen0

Community Trekker

Joined:

Dec 9, 2015

This may be a naïve question but I am confused about the 4PL equation used by jmp in curve fitting. I have a dataset of two variable and I choose: analyze>modeling>nonlinear

Then I plot the data and the red fit-curve button gives the option: Sigmoid curves>logistic curves>Fit logistic 4P

Now my question is; the equation used by jmp is f(x).

1 ACCEPTED SOLUTION

Accepted Solutions
michael_jmp

Staff

Joined:

Jun 23, 2011

Solution

Yes, the parameterization you give is different than the one used in JMP. The following is a reasonable discussion of the two parameterizations. JMP uses the one that it does to avoid having a parameter in an exponent, which makes estimation more difficult.

http://stats.stackexchange.com/questions/9298/reconciling-different-parameterizations-of-the-same-no...

-Michael

Michael Crotty
Sr Statistical Writer
JMP Development
6 REPLIES
r_o_pedersen0

Community Trekker

Joined:

Dec 9, 2015

It seems when I posted half my post was deleted??

...now my question is; the equation used by jmp is f(x)

Jeff_Perkinson

Community Manager

Joined:

Jun 23, 2011

Sorry you had trouble with your post. It seems though that there's still something missing.

Can you clarify your question?

-Jeff

-Jeff
r_o_pedersen0

Community Trekker

Joined:

Dec 9, 2015

Yes it seems everything after the equals sign gets deleted. Thanks for replying. My question concerns the logistic 4P fit in sasjmp. The equation used for this is not the same as for 4 parameter logistic fits used in various applications software I am used to. Is it just not the same model or is there some mathematical point I am missing? Usually 4PL is described as y equals ((A-D)/(1+((x/C)))+D

michael_jmp

Staff

Joined:

Jun 23, 2011

Solution

Yes, the parameterization you give is different than the one used in JMP. The following is a reasonable discussion of the two parameterizations. JMP uses the one that it does to avoid having a parameter in an exponent, which makes estimation more difficult.

http://stats.stackexchange.com/questions/9298/reconciling-different-parameterizations-of-the-same-no...

-Michael

Michael Crotty
Sr Statistical Writer
JMP Development
billw_jmp

Staff

Joined:

Jul 2, 2014

Hello,

I looked through the Nonlinear Model Library and I found the the 4P logistic model you describe is called the Rodbard Model (4P) in JMP.

10616_pastedImage_0.png

Best,

Bill

r_o_pedersen0

Community Trekker

Joined:

Dec 9, 2015

Thanks a lot. Those are great answers.