+ Reply to Thread
Results 1 to 3 of 3

Thread: Approximating the probability of an event occuring from the output of a Poisson GLM

  1. #1
    Points: 24, Level: 1
    Level completed: 47%, Points required for next Level: 26

    Posts
    2
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Cool Approximating the probability of an event occuring from the output of a Poisson GLM




    Hi all,

    I have some data that includes the output of a Poisson Log Link model, that aims to predict how many times an event will occur over the period of a year. I would like to approximate the probability of at least one event occurring from this output, though I am not sure how to go about it. I can get hold of the mean/base of the Poisson model very easily, which I'm guessing would be needed in any calculation.

    Any help would be massively appreciated.

    Thanks,

    Skinicod.

  2. #2
    TS Contributor
    Points: 12,287, Level: 72
    Level completed: 60%, Points required for next Level: 163
    rogojel's Avatar
    Location
    I work in Europe, live in Hungary
    Posts
    1,471
    Thanks
    160
    Thanked 332 Times in 312 Posts

    Re: Approximating the probability of an event occuring from the output of a Poisson G

    hi,
    maybe I misunderstand the problem but the answer seems to be P=1-P(0) where P(0) is exp(-lambda) lambda beeing the mean.

    regards

  3. #3
    Points: 24, Level: 1
    Level completed: 47%, Points required for next Level: 26

    Posts
    2
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: Approximating the probability of an event occuring from the output of a Poisson G


    I think I may not have explained very well, so let me put the question in context.

    A Poisson Log Link model has been built to predict the number of items that a customer may buy if marketed to.

    For each customer that has been marketed to, I have been provided with the models predicted frequency for the number of items that they will buy. However, I am more interested in answering the question - will they buy anything at all. If I were building a model to predict this, I would use a logistic regression, however all I have is the output of the frequency model.

    I am therefore trying to find a way to approximate one from the other. Is this possible?

+ Reply to Thread

           




Tags for this 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