Twin Cities JMP User Group Meeting on Friday, February 19 @ Noon Central

anspen — Thu, 11 Feb 2021 21:05:16 GMT

The next Twin Cities JMP User Group Meeting is taking place on Friday, February 19 at Noon Central. The agenda:

Seagate – Orthogonal Regression
3M – Split plot DOE
Ecolab - Graph Builder leading practices including how to add images to visualizations

Re: Twin Cities JMP User Group Meeting on Friday, February 19 @ Noon Central

ted_ellefson — Mon, 22 Feb 2021 14:47:01 GMT

Presentation from the Feb 19, 2021 event on Orthogonal Regression. We have successfully used this approach for gage to gage correlations.

Why would we want to do Orthogonal Regression?
What are the computational differences between ordinary least squares (OLS) and Orthogonal Regression
Perform an Orthogonal Regression study using JMP
A method originally attributed to W. Edwards Deming takes into account the error in both variables.
The result is completely invertible! It does not matter which is the “x” variable and which is the “y” variable.
The method is referred to as “orthogonal regression” (regression when both input and output have variability).
Creates a simple linear relationship only (no higher order terms): x2 = bx1 + a
Requires only a knowledge (or estimate) of the ratio of the error in y to the error in x.

Please feel free to contact me with any comments or questions.

Many thanks to those who contributed to this presentation over the years: Lyle Dockendorf, Joe Liu, Jim Gillard, Hugh Quinn, and Brenda Scott.

Orthogonal Regression uses an approach that minimizes the sum of the squared perpendicular differences
Requires that you specify the ratio of the variance of the error in X and Y