-1

When using the glm function: which is the best distribution for multinomial regression --- is it "binomial"? If the glm function doesn't have an appropriate distribution for multinomial regression, which r package has a better function? Thank you!

Health Policy Researcher
  • 11
  • 4
  • In general it's better to frame questions as "how do I do X?" rather than "where do I find a package that does X?", since the latter is considered a "software recommendation" question and thus off-topic for Stack Overflow (even if the *answer* turns out to be "get package Y") – Ben Bolker Apr 24 '21 at 20:23
  • Thank you! I am pretty new to stack overflow so I didn't realize this. – Health Policy Researcher Apr 26 '21 at 21:32

1 Answers1

-1

You may try the nnet::multinom() You can find the documentation at here.

David Z
  • 6,641
  • 11
  • 50
  • 101
  • I think the answer would be improved if you explained that `glm()` *can't* fit multinomial models and linked to https://stackoverflow.com/questions/42114194/can-multinomial-models-be-estimated-using-generalized-linear-model ... – Ben Bolker Apr 24 '21 at 20:22
  • Thank you @BenBolker. Your comment was very helpful! – Health Policy Researcher Apr 26 '21 at 21:34