Latent Dirichlet allocation

Latent Dirichlet allocation

From Wikipedia, the free encyclopedia

In statistics, latent Dirichlet allocation (LDA) is a generative model that allows sets of observations to be explained by unobserved groups that explain why some parts of the data are similar. For example, if observations are words collected into documents, it posits that each document is a mixture of a small number of topics and that each word's creation is attributable to one of the document's topics. LDA is an example of a topic model and was first presented as a graphical model for topic discovery by David Blei, Andrew Ng, and Michael Jordan in 2002.^[1]

来源：http://en.wikipedia.org/wiki/Latent_Dirichlet_allocation

重点介绍了相关的背景知识，比如马尔科夫链（Markov Chain）、条件独立以及判断条件独立的方法，包括马尔科夫毯（Markov Blanket）、贝叶斯球（Bayes Ball）等

重中之重中MCMC和Gibbs Sampling

并对其他几种主题模型进行了简单介绍，包括Author Model, AT（Author-Topic） Model, ACT(Author-Conference-Topic) Model等

希望大家明白Gibbs Sampling方法对inference来是多么简单的一件事情，从而导致非常简单的算法，相对于变分推导（Variational Inference）

现将今天的PPT贴在此处（ LDA.pdf），希望同道者一块探讨。

来源：http://blog.sciencenet.cn/home.php?mod=space&uid=611051&do=blog&id=518407