Validation of DoE (Design of Experiements) Model - Minitab 15

B

boucyr

Hello,
I am just starting to work with minitab 15.
I understand how to proceed ascreening (factorial design...) or an optimization (with RSM) but I do not understand how I could validate my model.
I ve used for practice datas which required Log-transformation but when I perform their analysis using Box-Behnken design for example with a quadratic model, I was not able to see that I need to perform this transformation, I was not able to get Q-square...
Could someone help me ?
Thanks !!!
 

Ajit Basrur

Leader
Admin
Hello,
I am just starting to work with minitab 15.
I understand how to proceed ascreening (factorial design...) or an optimization (with RSM) but I do not understand how I could validate my model.
I ve used for practice datas which required Log-transformation but when I perform their analysis using Box-Behnken design for example with a quadratic model, I was not able to see that I need to perform this transformation, I was not able to get Q-square...
Could someone help me ?
Thanks !!!

Welcome to the Cove, boucyr :bigwave:

Does anyone have answer to boucyr's question ? Thanks in advance
 
A

AdamP

A bit confused here as well. The original post cites a few issues - primarily wanting to validate a DOE model, but mentions 'practice' data. In my experience, you need to run it to generate real data based on the factor/level settings to confirm, or validate the model. Also, how you make the leap from screening to optimization and then bring in Box-Behnken is not clear. There are other optimization designs besides RSM - and Box Behnken can be used in that capacity - is that what you're trying to do?

Regarding the quadratic function - sounds more like a stats issue, but is that also the transfer function you're using as the experimental basis/platform?

If you can explain the question more clearly I'm sure there are folks here who can help guide you.
 
Top Bottom