Why is Matlab's control chart using standard error to calculate control limits?
3 views (last 30 days)
Show older comments
I am trying to generate a 'c' control chart with the following code:
[stats, plotdata] = controlchart(cData, 'chart', 'c');
Which generates the figure shown below and returns the associated statistics. The mean is calculated correctly as 1,107, but the LCL should be 0 and the UCL should be 3,363. The standard deviation of my data set is actually 751 and then the CLs should be calculated from this, however, Matlab looks to be using the standard error, which is calculated as 33 (thus the LCL is 1007 and the UCL is 1207 (mean +/- 3*se) as shown in the figure). So my question is how can I get the controlchart function to correctly calculate the standard deviation? Note that I tried doing this as well, to no avail (in fact it didn't change anything):
[stats, plotdata] = controlchart(complaintsPerMonth, 'chart', 'c', 'sigma', 751);
Thanks in advance for any help!
2 Comments
the cyclist
on 14 Nov 2014
I'm no expert, but the Wikipedia page on control charts states in more than one place that the chart uses standard errors.
KAE
on 3 Aug 2020
No expert either, but this does not look like a normally distributed process with a mean value that stays pretty constant over time. Also it seems like the input argument 'sigma' can be your choice of method for estimating the upper and lower bounds. Look at the controlchart.m code,
edit controlchart.m
and you will see the choices.
Answers (0)
See Also
Products
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!