My customer asked me to test a few test of data for independent data. He told me to use Run Chart and check p value. However i am unable to determine how this is done.
Basically what we are trying to achieve is take data sets for a particular characteritsics from many machines,
Our intent is to check the following, We should test for normality, independency and hypothesis testing for mean and equal variances. I am able to do all the test except independency. Any help to solve this will be useful.
There is an JMP Addin in the File Exchange that will calculate the Run's Test you are looking for.
Just download and install the Adding, and you can then run it.
Sometimes when I hear about independence of runs, the concern has to do with autocorrelation. That's easy to calculate, but even easier to turn on in the Distribution Platform, (summary statistics, red triangle, Customize summary statistics, turn on autocorrelation.
To pile onto my colleague @Byron_JMP you may also want to take a look at the Specialized Modeling -> Time Series platform if indeed your data is in time series form by examining autocorrelation at lag time t.
I'm not big on using any single 'metric' to evaluate autocorrelation within time series data when one has the rich set of data visualization options within JMP. Within the Analyze -> Specialized Modeling -> Time Series platform report the Time Series Basic Diagnostics autocorrelation and partial autocorrelation plots provide lots of useful information such as whether or not autocorrelation is likely to exist...and then if it's apparent, at what lag? A great way to see how the platform works is to run the embedded analysis script in the JMP data table, 'Seriesg' (in the JMP Sample Data Directory).
Here's some additional information within the JMP online documentation: