I have a relatively small data consisting of 10 -15 experiments, but to run any statistical process control, one would need a large sample to capture enough variability in the data and be able to compute Cpk values. I tried bootstrapping but the confidence interval of mean is very narrow and not very useful. Running more experiments is not possible as it is quite expensive both by time and money. What are the options that one could use in such a case to statistically validate the results?
Farid Hassanvand. Thanks for answering. Could you please provide some more information on using the fuzzy methods in SQC / SPC. Or perhaps, if you could share some paper ...
I think the best way for you is the bootstrap technique. There are various form of confidence interval for bootstrap method, you can use them.
Meanwhile, for using fuzzy method your data must be inaccurate or imprecise.
- Badges
- Science topic
- Similar topics
- Biological Science
- Kinesiology
- Running