Logistic regression prediction formula

Discussion in 'microsoft.public.sqlserver.datamining' started by goldin, Apr 26, 2007.

  1. goldin

    goldin Guest

    In our application, we are trying to calculate predictions from logistic regression after we extract from the mining model contents the estimated coeficients. We understand that we have to "standrize" the input variables before we multiply them by the coeficients but we did not found the correct standartization formula. We appreciate help on that subject.
     
    goldin, Apr 26, 2007
    #1
    1. Advertisements

  2. goldin

    Dejan Sarka Guest

    In our application, we are trying to calculate predictions from logistic regression after we extract from the mining model contents the estimated coeficients. We understand that we have to "standrize" the input variables before we multiply them by the coeficients but we did not found the correct standartization formula. We appreciate help on that subject.

    I think it is Z-score normalization: v = (x - µ) / σ
     
    Dejan Sarka, Apr 26, 2007
    #2
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.