60985_en-US

liliconcom

Scripting - Best Fit for Process Capability

Hello, 

 

I have a function inspired from this Post that seeks to do the process capability in a generalized way for my needs. The one issue I am having is setting this so that it does not assume Normal distribution. I want this to look at each column and choose the best fit option for the analysis. 

 

Here is the function I have more or less (I have cut out simple changing names of columns and such). Is there a simple way of adding to this to accomplish using best fit instead of assuming normal? 

 

ProcessFunc = Function( {TheList}, 
 TheVars = ":Name(\!"" || TheList[1] || "\!")";
 For( i = 1, i <= nitems(Rcols), i++,
  TheVars = TheVars || "," || ":Name(\!"" || TheList[i] || "\!")"
 );
 pcVar=Eval(Parse("dt5 << Process Capability( 
  Process Variables( "|| TheVars|| ")) 
  );"));
  pcVar << Overall Sigma Summary Report;
  pcVar << Save Spec Limits as Column Properties;
);

Re: Scripting - Best Fit for Process Capability

Awesome, I was able to get this working using this idea and slightly changing it. Thank you so much!  

Re: Scripting - Best Fit for Process Capability

Personally, I'd recommend the non-parametric option as being more robust than trying to automate a best fit.  I think you could accomplish what you're looking for by appending "&Dist( Best Fit)" to each column in your list using a simple loop.  Try this below your definition of TheVars:

for(i = 1, i<=N Items(TheVars), i++,
 TheVars[i] = parse(char(TheVars[i])||"&Dist(Best Fit)");
);