[Topic-models] Why does alpha serve as a scalar in blei's code

Sinead Williamson saw56 at cam.ac.uk
Sat Jan 15 08:52:42 EST 2011

Hi Peng,

You might find the following paper of interest:

Hanna Wallach, David Mimno and Andrew McCallum. "Rethinking LDA: Why
Priors Matter." In Proceedings of the 23rd Annual Conference on Neural
Information Processing Systems, 2009.

Best wishes,


On Sat, Jan 15, 2011 at 1:45 PM, PENG Peng <ppeng at ust.hk> wrote:
> Hi everyone,
>   I do not quite understand why Blei's LDA code considers the parameter
> alpha as a scalar so that the learned dirichlet distribution is a
> K-dimensional symmetric Dirichlet. In my understanding, the elements in
> vector alpha could be different. Can anyone explain for me? Thank you
> so much.
> Regards,
> Peng
> _______________________________________________
> Topic-models mailing list
> Topic-models at lists.cs.princeton.edu
> https://lists.cs.princeton.edu/mailman/listinfo/topic-models

Sinead Williamson

Computational and Biological Learning
Department of Engineering
University of Cambridge

More information about the Topic-models mailing list