MCMC Method And Its Application In The Text Topic Modeling

Weibo because of its Openness,low threshold,the terminal expansion,concise features,information sharing has become an important platform for dissemination and access of popular news and current affairs.But now microblogging mass text basically fast-paced short text data,the situation is now,microblogging text on the Internet as ubiquitous as text,every day in huge data traffic generated all the time kept.Needless to say,reasonable and effective analysis of these data can bring us great value.In this article,we highlighted MCMC method and LDA topic model as well as a special MCMC method: Gibbs Sampling Sampling Algorithm.LDA model relates to the Dirichlet distribution and several different text modeling.Because LDA topic model involves more pre-theoretical model,Gibbs sampling algorithm is more theoretical basis of statistical algorithms.In this paper,for the sake of clearer explanation of the above we have carried out a more thorough and detailed description and interpretation.In order to test the practical effect of LDA Model and Gibbs Sampling algorithm,and think about how to apply it to practice,the paper design and completion of the "microblogging users interested in discovering the system." The system can be gathered with a user microblogging together with the LDA model for relating to mining,digging out each topic is the user's interests and hobbies.