Gaussian LDA (2): Gaussian LDA Introduction

Latent Dirichlet Allocation (LDA) is a thematic model that enables the modeling of text and the distribution of the subject matter of the document. But each of the topics that LDA gets is a multi-item distribution on terms that is very sparse. In order to describe the semantic coherence better, some researchers have proposed Gaussian LDA, this paper introduces the model briefly.

Reprint Please specify source: http://blog.csdn.net/u011414416/article/details/51188483

This article mainly introduces the following work of ACL2015:
Rajarshi Das, Manzil Zaheer, and Chris Dyer. Gaus-sian LDA for topic models with Word embeddings. In Proceedings of ACL 2015.

In addition, the main reference is Rickjin's "Lda Math Gossip"
and parameter estimation for text analysis this technical report

Gaussian LDA (2): Gaussian LDA Introduction