Is it possible to have multiple (at least 3) dependent variables in a single classification model. I know this can be accomplished in a regression model but I need to perform this with a classification model.

My three DVs are correlated so I do not want to simply write three seperate classification models. Does anyone know of a way to do this in either R or python's scikit?

  • $\begingroup$ Neural networks would allows for such things in a pretty straightforward way. $\endgroup$ May 13, 2014 at 16:02
  • 1
    $\begingroup$ Do you have any examples to show this? I have looked extensively and have found no such example. $\endgroup$
    – mike1886
    May 13, 2014 at 17:50
  • $\begingroup$ In a neural network setting, I would add new output nodes for every extra DV. You can reuse the rest of the architecture (e.g. all hidden layers, which constitute the main part of learning). $\endgroup$ May 13, 2014 at 19:58


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.