topic Re: Standard Error of Regression for a Fit Model in Discussions

Standard Error of Regression for a Fit Model

Shujinko — Fri, 09 Jun 2023 00:41:57 GMT

I used the Fit Model platform to generate a regression model for my data. I would like to find the standard error of regression (S) so I can roughly estimate if my regression model is precise enough to my needs, e.g. I need my model precision to be within +/- 5% of the actual values to be useful.

I know that Minitab offers this value S along with the R^2, and I know that JMP offers standard error estimates for individual parameters.

1. Does JMP indicate what S would be for the model and if so am I just missing it?

2. If it doesn't offer it, could I go about calculating it from the existing values JMP gives?

I have attached an image of my analysis.

Re: Standard Error of Regression for a Fit Model

statman — Tue, 16 Nov 2021 02:46:21 GMT

Sorry, I may be confused by your question, but do you mean RMSE? If so, yes this is part of the Summary of Fit along with R-Square and R-Square Adj. Whether or not this is useful for your application is a different question.

Re: Standard Error of Regression for a Fit Model

Dan_Obermiller — Tue, 16 Nov 2021 03:28:27 GMT

@statman is correct. The Minitab "standard error of regression" is also known as the Root Mean Square Error (RMSE). It is in the Summary of Fit report that JMP provides, right near the RSquare values.