Activity

If it is possible for you to post your data that would be quite helpful. As already noted by others this may well be a case of inadequate measurement discrimination (aka “chunky data”), stability or bimodality. 5 years, 3 months ago

John, StdDev for nonnormal Cpk uses MRbar/1.128 of the zscores.
7 years, 6 months ago 
I do need to add that SigmaXL only supports nonnormal capability for individuals not subgroups, so that may be a limiting factor. 7 years, 6 months ago

Note that Minitab does calculate both Ppk and Cpk when using the BoxCox transformation. It does not do so for the Johnson transformation or nonnormal distributions.
An important assumption in this discussion is the data are inherently nonnormal, not nonnormal due to outliers, bimodality or a “chunky data” measurement problem.
Given that…[Read more]

SigmaXL includes Cpk for nonnormal capability using a normalized within standard deviation. Statgraphics also does this. 7 years, 6 months ago

To confirm Robert’s comment, using SigmaXL’s robust sample size template for One Way ANOVA (based on Monte Carlo simulation), a small sample size of n=9 per group will give you good robustness against severe skewness (Skew = 2).
Note, if you are doing posthoc tests with Welch’s ANOVA, you should not use Tukey, but rather GamesHowell, which is…[Read more]

John Noguera commented on the post, Optimize Attribute Responses Using Design of Experiments 8 years, 4 months ago
Maxim,
Thank you for the detailed response comparing FTP to BLR.
In order to optimize the Minitab Event Probability model, you could set up the equation in Excel and then use Solver or DiscoverSim to perform the […]

John Noguera commented on the post, Optimize Attribute Responses Using Design of Experiments 8 years, 4 months ago
Maxim, Michael,
Thank you for your comment. There is definitely a trade off for BLR on statistical accuracy versus difficulty of interpretation. I would be very interested to hear if you come up with a Rule […]

John Noguera commented on the post, Optimize Attribute Responses Using Design of Experiments 8 years, 5 months ago
Thank you for the interesting article. However, this approach does not take into account the sample size of your replicates. There is a statistical difference between 7/10 and 700/1000!
The way to do that (and […]

I wonder why the physicists didn’t include a 1.5 sigma shift in their announcement :)
It’s refreshing to see 5 sigma mean 5 sigma!
8 years, 8 months ago 
Hi Eric,
Thank you for the compliment! I can’t take any credit for the ARMA tool mentioned above.
We will be adding full BoxJenkins ARIMA in our next release. 8 years, 8 months ago

Figured out the problem. The link works with IE but not Chrome. 8 years, 8 months ago

The ARMA tool should do the trick if you can get it. Try contacting the author on the web site I gave you.
It has a differencing function, so you can specify p,d,q.
The Correlogram/ACF is also mentioned as a separate tool available for download.

Somewhat limited but try this free Excel addin for ARMA Modeling:
http://www.webreg.de/arma_addin.html
Update – link appears to be broken.
Alternatively:
Set up a manual EWMA in Excel, with the Lambda value as a separate input cell.
Create an output cell that is the sum of residual^2.
Use Solver to minimize the output, by…[Read more]

John Noguera and Ghisoni are now friends 9 years, 1 month ago

John Noguera and RAMDASI are now friends 9 years, 1 month ago

John Noguera changed their profile picture 9 years, 2 months ago

John Noguera became a registered member 9 years, 2 months ago