0
$\begingroup$

I am trying to create a code in Python to select orthogonal profiles given some attributes and levels. For eg:

{'Pressure':[40,55,70,80,90],
'Temperature':[290, 320, 350],
'Flow rate':[0.2,0.4,0.5,0.6,0.7],
'Time':[5,11]}

If I need to a conjoint analysis using all possible profiles, it would become 535*2=150 profiles, which is too much. I need to reduce the number of profiles so that they are orthogonal. Is there any implementation in Python that I can use or any guidance on how I would go about building this would be highly appreciated.

Thanks!

$\endgroup$

1 Answer 1

0
$\begingroup$

Sorry, no experience with python, but fairly easy to do in R, even for absolute beginners:

# if you do not have it yet, install via install.packages("radiant")
library(radiant) 
doe(c("Pressure;40;55;70;80;90",
      "Temperature;290; 320; 350",
      "Flow rate;0.2;0.4;0.5;0.6;0.7",
      "Time;5;11"),
    trials = 50)

They also have Shiny-App: https://vnijs.shinyapps.io/radiant/?SSUID=3dcff236d7

Go to Design, then to doe, put in your Design Factors:

Pressure;40;55;70;80;90
Temperature;290; 320; 350
Flow rate;0.2;0.4;0.5;0.6;0.7
Time;5;11

Set trial to the number of profiles you can afford. That's it. Note that you will not get a perfectly orthogonal design, but it will usually be ok as long as the trial number is not too small.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.