Logistic distribution RRS feed

  • Question

  • I'm trying to create a variable with a logistic distribution and I read on another forum post (http://social.microsoft.com/Forums/en-US/daa215eb-3a08-44c1-b1d7-f7433193af4b/new-distributions-migrated-from-communityresearchmicrosoftcom?forum=infer.net) that this can be done through the Gamma distribution.

    From Tom's post: 'logistic = log(F(1,1)) = log(Gamma(1) with Gamma(1) rate)'

    So would I specify the Variable like so?:

    Variable<double> logistic_var = Variable.Log(Variable.GammaFromShapeAndRate(1, Variable.GammaFromShapeAndRate(1, 1)));

    If so, which components of that line refer to the location and scale parameters of the logistic distribution? Can Infer.net learn both of these parameters?

    Thanks very much for all the help,


    Sunday, August 24, 2014 5:05 PM


  • There are no location and scale parameters in that line.  You would need to add them.  If m and s are Variables, then m+s*Variable.Log(...) gives you a logistic distribution with location m and scale s.
    • Marked as answer by MarkG87 Wednesday, September 3, 2014 3:02 PM
    Tuesday, August 26, 2014 10:33 AM