Well Tested Latent Dirichlet Allocation in Python

I need to integrate a python LDA module in my tool. It seems there are number of LDA oracles out there. I am wondering if folks here have tested any of the tools and can point me to a well tested one that I can just take and integrate with my other libraries.

python lda

asked Feb 27 '11 at 13:57

Mark Alen
1323●23●41●46

2 Answers:

oldest newest most voted

4	I'd use Matt Hoffman's implementation of VB and online VB for LDA, as it works well, is well debugged, and is a lot faster than the lda code I have online (which is poorly maintained). answered Feb 27 '11 at 15:02 Alexandre Passos ♦ 25541●54●278●421

gensim has a fast Python implementation of LDA. It can process input larger than RAM, and it can run distributed over a cluster of computers (if you have one).

The implementation is based on cleaned-up Hoffman's onlineldavb code suggested by Alexandre Passos. Processing the English Wikipedia with gensim takes 11 hours on a laptop.

answered Mar 24 '11 at 02:50

Radim
386●7●9

edited Mar 24 '11 at 09:22

Subscription:

Once you sign in you will be able to subscribe for any updates here

Tags:

lda ×114
python ×97

Asked: Feb 27 '11 at 13:57

Seen: 8,892 times

Last updated: Mar 24 '11 at 09:22

Well Tested Latent Dirichlet Allocation in Python

Subscription:

Related questions