How to form Clusters and subsets?

sanqub

Dear JMP Users,

 

Apologies, question might be misleading.

I am studying a manuafacturing process and the output is categorical, Good and bad. Sometimes in process a defect occurs continuously for 10 or more times until a good part is found. My aim is to find such consecutive parts wherein the defect has repeated number of times. I will later use this subset to study the similarities if a particular variable range(or group of variables) is causing this defect. This is advice from my Professor for different kind of analysis. I would request for your suggestions to form subset in JMP.

I guess using Clusters/PCA will not necessarily give cluster of consecutive defects. Any other option I can use to get the required subsets.

 

Thanks

 

San  

 

 

 

 

Peter_Bartell

Since viewing the data in time series sequence is important you may want to consider using the Process Screening platform to help answer the 'When does something happen?' part of your problem. Then you could just visually inspect the part disposition (good or bad) and process variables for coicindences in time when things appear to change in phase.

 

Also, since this sounds like an academic problem...usually there are certain analytic methods that are part of the pedagogical experience that are in play. What does your professor suggest?