DOC Note how ariddell/lda differs from sckit-learn's LDA (scikit-learn#5553)

ariddell · paulha · commit 334ffb03fc4b · 2017-08-18T21:52:38.000-07:00
Fixes scikit-learn#5529
diff --git a/doc/related_projects.rst b/doc/related_projects.rst
@@ -84,8 +84,14 @@ and tasks.
   ensembles) trained by sklearn. Useful for latency-sensitive production
   environments.
 
-- `lda <https://github.com/ariddell/lda/>`_: Fast implementation of Latent
-  Dirichlet Allocation in Cython.
+- `lda <https://github.com/ariddell/lda/>`_: Fast implementation of latent
+  Dirichlet allocation in Cython which uses `Gibbs sampling
+  <https://en.wikipedia.org/wiki/Gibbs_sampling>`_ to sample from the true
+  posterior distribution. (scikit-learn's
+  :class:`sklearn.decomposition.LatentDirichletAllocation` implementation uses
+  `variational inference
+  <https://en.wikipedia.org/wiki/Variational_Bayesian_methods>`_ to sample from
+  a tractable approximation of a topic model's posterior distribution.)
 
 - `Sparse Filtering <https://github.com/jmetzen/sparse-filtering>`_
   Unsupervised feature learning based on sparse-filtering
