Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

A Logistic Regression Interpretation

  1. Mar 17, 2017 #1
    I was trying to find an easy interpretation of the predicted probabilities of a logistic regression model, when one of my coworkers claimed that the logistic regression model is a likelihood.

    Now, I know that maximum likelihood estimation is used to estimate the parameters, but I didn't think of the model as a likelihood.

    The model is E(Y|X)_hat = exp(XBeta_hat)/(1+exp(XBeta_hat)).

    Is the above function a likelihood function?
    Last edited: Mar 17, 2017
  2. jcsd
  3. Mar 18, 2017 #2
    I'm going to answer my own question. No, the predicted probability is not a likelihood.

    The likelihood is the probability density function, as a function of the data. That is L(p|y) = f(y|p), for a fixed y. The likelihood is telling you how likely p is for a specific value of p, given the data y.

    Since the predicted probability does not include all the data (it doesn't include y), you cannot conclude it is a likelihood.
    Last edited: Mar 18, 2017
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted