cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
JMP is taking Discovery online, April 16 and 18. Register today and join us for interactive sessions featuring popular presentation topics, networking, and discussions with the experts.
Choose Language Hide Translation Bar
miguello
Level VI

Need help understanding parameters and table variables

Hi!

 

I have a practical problem that I think I could solve using some kind of parameter or table variable.

Let's say I have multiple processes, each having a row. Which are characterized by few parameters, A, B, C... Those are columns.

Let's say for each parameter I have a threshold after which it is considered risky. So I created columns bit-A, bit-B etc. If parameter is grater than some number, I give it bit "1", if it's below - I give it bit "0". Then I do logical  AND for all the bits and if it's "1" (all parameters are in "risky" zone) I call the whole process "risky". Let's say final risky bit is saved in column bit-Risky.

 

So, I'm sure about thresholds on A and C, but I'm not sure about threshold on B. So I want to see how number of risky processes depends on threshold for B (let's call it dB).

 

So, how do I plot Sum(bit-Risky) vs. dB?

 

1 ACCEPTED SOLUTION

Accepted Solutions
gzmorgan0
Super User (Alumni)

Re: Need help understanding parameters and table variables

@miguello There might be a method to do this with linked profilers, but I think you might want to do this with a script.

 

Attached is a table, my representation of what you described. The table has a script to aggregate (find the sum) for various scenarios, various thresholds and store them in a table, then plot the Sum vs. dB.

 image.png

This could be done with an interactive window and made quite fancy.  But this is just a simple script with hard-coded names, and thershold ranges.

 

In the past, I have experienced similar requests for results of complex simulators, like semiconductor simulators.  Instead of counting, I think there is more insight by doing some analysis on the relationships of pass/fail.  Also attached to the table is a parallel plot (see below) script that can provide some insight to the variables.  Here a Local filter could be added and ranges of A, B and C can be selected.

image.png

View solution in original post

2 REPLIES 2
gzmorgan0
Super User (Alumni)

Re: Need help understanding parameters and table variables

@miguello There might be a method to do this with linked profilers, but I think you might want to do this with a script.

 

Attached is a table, my representation of what you described. The table has a script to aggregate (find the sum) for various scenarios, various thresholds and store them in a table, then plot the Sum vs. dB.

 image.png

This could be done with an interactive window and made quite fancy.  But this is just a simple script with hard-coded names, and thershold ranges.

 

In the past, I have experienced similar requests for results of complex simulators, like semiconductor simulators.  Instead of counting, I think there is more insight by doing some analysis on the relationships of pass/fail.  Also attached to the table is a parallel plot (see below) script that can provide some insight to the variables.  Here a Local filter could be added and ranges of A, B and C can be selected.

image.png

miguello
Level VI

Re: Need help understanding parameters and table variables

Thanks! This definitely works!
However, my specific problem is a little bit more complex than I initially described.
Each process has around hyndred test locations. So, just getting a sum of bits wouldn't do it.
Plus the conditions on thresholds might be complex as well - like, range, text comparison etc.
The way I eventually did it - I was just iterating through thresholds in the script. Luckily, all the bit columns were updated on threshold change. All I needed to do is to get a summary table and then I could get my risky bits as matrix and sum them - as you suggested.