"All possible models" function in JMP 9.0

[nangell]

Hello,

I recently used the "all posible models" function in the stepwise platform of JMP 9.0.  I am wondering what process JMP is using to compute the values?  Is JMP performing backwards and/ or forwards stepwise regression or is it using another procedure?  Also, if JMP is using stepwise regression what are the criteria that it is using?

Thanks for your help.

natalie

[mewing]

Natalie,

I believe the 'all possible models' option in the stepwise platform simply considers all combinations of factors up to a set number of factors being in the model.  That is, if you have 3 x's, call them A, B and C then JMP would fit all the following models:

A

B

C

A+B

A+C

B+C

A+B+C

For each of the models it would calculate fit and then list them so you could identify the best.  So, instead of using a logic based stepwise system, it's a 'brute force' type of model fitting.