topic Re: Multivariate model driven control charts with historical data ending at certain row in Discussions

Multivariate model driven control charts with historical data ending at certain row

Chitranshu — Fri, 09 Jun 2023 14:42:03 GMT

When ever I am relaunching analysis after specifying that the historical data ends at certain row. I am getting error.

Re: Multivariate model driven control charts with historical data ending at certain row

yuichi_katsumur — Wed, 10 May 2023 04:37:02 GMT

How did you launch a MDMCC? If you ran PCA or PLS and then saved the principal component formulas and launched the Model Driven Multivariate Control Chart (MDMCC) launch window, you must use the same number of historical data rows as used in PCA or PLS. Please see the note on the following page. https://www.jmp.com/support/help/en/17.1/#page/jmp/launch-the-model-driven-multivariate-control-chart-platform.shtml#

If the above procedure is not what you have done, could you tell us how to launch MDMCC in more detail?

Re: Multivariate model driven control charts with historical data ending at certain row

Chitranshu — Wed, 10 May 2023 20:40:41 GMT

Hi @yuichi_katsumur

I have taken my data eg: 10 years and after that I ran PCA. In PCA I applied filter to determine the stable period of operation (say 2 years). After this I launched MDMCC. But whenever I am making historical data end, I am getting same error,

Re: Multivariate model driven control charts with historical data ending at certain row

yuichi_katsumur — Thu, 11 May 2023 04:26:53 GMT

Hi @Chitranshu ,

It seems that JMP doesn't support the following procedure.
- using the local data filter in the PCA platform and then launching MDMCC from the red triangle.
This is a workaround.
- Launch PCA, then use local data filter and decide stable period.
- Hide and exclude unused rows using the Rows > Hide and Exclude menu.
- Choose Analyze > Quality and Process > Model Driven Multivariate Control Chart.
- Select Monitor the Process > Set Component from the red triangle.
Hope it helps.