Monte Carlo Simulation

[#1]

Hi hopefully someone can save me some time!

I have created models for 20+ responses with same inputs and want to do MC simulation for all responses.  I did it for the first one as follows:

1) from the Fit Least Squares window, red triangle on first response, Save Columns, Prediction Formula

Q1) do I have to do the same steps 20+ times for the other responses or is there a quick way to make the Prediction Formula column for all responses modelled?  I'd rather not use JSL I'd just like the equivalent of holding the control key down to make an action apply to all responses.

2) In Data Table, did Graph, Profiler, set the Prediction Formula column to be Y and the 4 inputs to be Noise factors, OK. Did Prediction Profiler, Simulator, set each of my inputs to Random and set mean and stdev. Ran simulation which looked good, saved table.

Q2) How do I hide the d(resp1)/d(input1) plots in the simulator graph, I only want to see the responses otherwise there are going to be 20+ responses times 4 inputs = 80+ rows of subplots which is not very helpful.

All advice gratefully received.

Justin

[#2 · Accepted solution]

Q1) If you fit all the models in the same report, instead of individually, you can hold the Ctrl-key down and use Save Columns -> Prediction Formula. Since they are all using the same inputs fitting them this way shouldn't be a problem. Just put them all in as Y columns in the Fit Model dialog.

You can find a joint Profiler at the bottom of the Fit Model report with multiple responses, so you don't even really have to save the prediction formula.

Q2) You don't need to put the inputs in as Noise Factors. That's what's getting you the derivative responses. Just enter all the prediction formulas as the Ys and you'll get what you're looking for.

I hope that helps!

-Jeff

[#3]

OK, I was pressing Control only for the last click (Prediction Formula) which didn't work but holding Control from red triangle onward does - thanks.