nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Spatial Wafer Clustering - Save Scoring Formula

Phil_Kay

Hi,

A customer in Semiconductor likes the spatial clustering tools but asks if it is possible to save a formula to score new wafers. This would tell them which cluster any new wafer belongs to. I can't find a way to do this. Is it possible?

Thanks,

Phil

Accepted answers

MikeD_Anderson

Normally, you can use the "Save Formula for Closest Cluster" but, for some reason, that option isn't available with spatial measures turned on. The next best thing to do would be to make the clusters save them back to the data table and build a predictive model using the cluster as the response. I think that's the only way to get where they would want to be in this case. If they've got Pro, a Bootstrap Forest with stratified validation would be the strategy I'd try first.

M

All comments

MikeD_Anderson

Phil_Kay

Thanks. I thought that was the case.
Cheers,
Phil

MikeD_Anderson

I also put it in as a defect with the JMP Beta mailing list. You might put the feedback in there so they have the details.

M

Phil_Kay

Thanks, Mike.

I suspect the difficulty is on the stacked format of the original data. I guess you would have to create a split version of the data in order to be able to have a scoring formula for an individual wafer.

Regards,

Phil