topic Re: Prediction Profiler's Simulator Question - How to changing the P-Value in Discussions

Prediction Profiler's Simulator Question - How to changing the P-Value

Radian2022 — Sat, 10 Jun 2023 23:46:56 GMT

Greetings,

I am trying to use the JMP's prediction profiler's Simulator to run a MonteCarlo simulation using a data-set that I have. After setting up the simulator and specifying the random factors and their corresponding SD values, I can see the simulation output - no problem there. However, I noticed that the simulator is applying +-2 Sigma (95% confidence interval) to the factors to run the montecarlo. What I'm really interested in is to do this simulation with +-6 Sigma variation in the factors and to see the outcome. I spent sometime on this but was not able to figure it out. Can anyone help me with this ? Your inputs are much appreciated.

Re: Prediction Profiler's Simulator Question - How to changing the P-Value

Emmanuel_Romeu — Fri, 08 Apr 2022 06:32:48 GMT

Hello, what about using 3 SD rather than 1 SD (not a perfect solution I have to admit).

Re: Prediction Profiler's Simulator Question - How to changing the P-Value

Radian2022 — Fri, 08 Apr 2022 16:18:10 GMT

Do you mean increasing (i.e., inflating) the numerical value of SD ? If so, I am afraid that'll skew the distribution, resulting in a pessimistic prediction.

Re: Prediction Profiler's Simulator Question - How to changing the P-Value

Mark_Bailey — Mon, 11 Apr 2022 18:16:40 GMT

You select the appropriate distribution to represent the expected variation in the factor settings. You do the same for the response (variation no associated with factor level changes). You define the distribution parameters (e.g., mean and standard deviation of the normal distribution). JMP is not using an interval such as plus or minus 2 SD or 6 SD.

I might have misunderstood your description or your question.