Statistica: Prediction of values in Design of Experiments

[Hannes Richter]

Hello,

 

I'm pretty new to Statistica. So I hope my question is pretty easy.

 

I did some engineering experiments with variation of different parameters with a full factor plan. Now I want to validate whether a semi factor plan (e.g.) would have led to similar results in order to maybe reduce the test scope in further investigations.

 

So is it possible to pedict values for not tested combinations of parameters Or how do I validate the application of non-full-factor plans for my eperiments

 

Best regards

Hannes

[Tomas Jurczyk]

Hello Hannes.

If I understand correctly, you have already conducted full factorial design and gather the data from experiment and the question is if you could simplify the design, is this proper understanding

For going to simpler design, I suppose an important question is if the interactions in the model are important (if you conducted the measurements already, you will have this info). If interactions are not important, design could be simpler without losing much information.

If you have data already measured for full factorial,you can do computations in the tool for simpler designs easily because you have data already available and after that you can compare results.

I am not sure, if I understood the question properly, so feel free to clarify if answer is not helping.

Best

Tomas