VIP plot in JMP PLS platform

[#1]

Hi,

I have a question reagarding the Variable Importance in the Projection plot in JMP.  In JMP, i only get one VIP plot with default cut-off line at 0.8. However, if we do PLS methods, shouldn't we have n of VIP plots (n is the component number) ? So I wonder why JMP only has one VIP plot? My guess is JMP use the first component in PLS?

Hope someone could give some insights about this. I checked the JMP online docs but didn't find any explanations.

Thank you,

Tao

[#2]

Basically, in PLS VIP relates to the model rather than the latent factors that appear within it, See the documentation, or, or more details, Discovering Partial Least Squares with JMP.

[#3]

Hi Ian,

Thank you for your reply! I will check the results. The reason I am asking this question is because when i use other statistical software packages like R, when i use PLS and VIP function, they tend to generate mutiple VIP plots and that's why I wonder what are the differences between the ones in JMP and R.

Tao