Beta Matrix

[rastash]

Hi authors,

Thank you very much for the very useful open source implementation of LDA via map/reduce. I have one question about the output. What does the beta matrix contain in your implementation? Typically, it is the log likelihood of terms in topics. However, I noticed that the HMapIDW output in your beta file contains both positive and negative numbers, which makes me wonder if it is actually some other transform rather than log likelhood?