900180_en-US

liliconcom

How I get best transformation using box cox transformation

In box cox plot, jmp not recommended any transformation, if lambda is 1 then no transformation required but if lambda is greater or smaller than 1 then which transformation we choose. JMP offer 4 options, refit with transform replace with transform, save best transformation, and sav specific transformation, which one is suitable.

Design of ExperimentsRe: How I get best transformation using box cox transformation

Hi @Keshav,

Welcome in the Community !
I think you can find the answer in the documentation.

To summarize, the first three options are different display/use/save of the best transformation found, whereas the last one enable you to manually change the lambda coefficient (so manually change the transformation):

• Refit will relaunch the modeling with the best transformation in another window,
• Replace will modify the model from the already opened window with the best transformation,
• Save best transformation will save a prediction formula column in your datable (with the best lamba coefficient found for the transformation),
• Save specific transformation enable you to specify the lambda coefficient and save this transformation prediction formula with this specific transformation in your datatable.

 

Hope this answer will clarify the available options,