+ Reply to Thread
Results 1 to 5 of 5

Thread: Optimal cut-off from ROC

  1. #1
    Points: 731, Level: 14
    Level completed: 31%, Points required for next Level: 69

    Posts
    11
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Optimal cut-off from ROC




    Hello,

    as you know, ROC curve is not depend of the cut-off. But how can I get an "optimal" cut-off from ROC curve?

    Example: I have a database which consist of 1 400 observations (clients) and 700 of them are "failed clients". When I conduct/perform a logistic regression I get coefficients/statistics/plots and so on. But when I want to depict a classification table which cut-off should I use? Which cut-off is optimal for decide which client is bad/good?

    Is it possible to find out this information from ROC curve? (please find attached )

    Thank you very much!

    Martin
    Attached Images  

  2. #2
    TS Contributor
    Points: 40,089, Level: 100
    Level completed: 0%, Points required for next Level: 0
    Awards:
    Downloads
    gianmarco's Avatar
    Location
    Italy
    Posts
    1,367
    Thanks
    232
    Thanked 301 Times in 225 Posts

    Re: Optimal cut-off from ROC

    Which program are you actually using?
    http://cainarchaeology.weebly.com/

  3. #3
    Points: 731, Level: 14
    Level completed: 31%, Points required for next Level: 69

    Posts
    11
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: Optimal cut-off from ROC

    I have used SAS

  4. #4
    TS Contributor
    Points: 40,089, Level: 100
    Level completed: 0%, Points required for next Level: 0
    Awards:
    Downloads
    gianmarco's Avatar
    Location
    Italy
    Posts
    1,367
    Thanks
    232
    Thanked 301 Times in 225 Posts

    Re: Optimal cut-off from ROC

    Unfortunately, I am not familiar with SAS. In R it would be rather simply.
    See this earlier thread

    hope this helps
    Gm
    http://cainarchaeology.weebly.com/

  5. #5
    Omega Contributor
    Points: 38,303, Level: 100
    Level completed: 0%, Points required for next Level: 0
    hlsmith's Avatar
    Location
    Not Ames, IA
    Posts
    6,993
    Thanks
    397
    Thanked 1,185 Times in 1,146 Posts

    Re: Optimal cut-off from ROC


    Depends on how you define "optimal trade off". If you just want the best blend of false positives and false negatives, the Youden Index can work. Let us know if you have not solved your question.


    You can use the ctable option in SAS PROC LOGISTIC statement and use those data to find the best answer. I have also, noticed regression trees also seem to generate the same cutoff as ROC based approaches.
    Stop cowardice, ban guns!

+ Reply to Thread

           




Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts






Advertise on Talk Stats