search

ryanneph / GibbsSampler_BagOfWords

Very simple implementation of gibbs sampling on the classical unsupervised bag-of-words bayesian clustering problem for natural language document topics.
0 stars 0 forks source link
bayesian-inference gibbs-sampling natural-language-processing

readme

Gibbs Sampler for Bag-Of-Words Topic Inference

Gibbs sampler for the 2 class bag-of-words topic clustering problem applied to document word count datasets downloaded from http://archive.ics.uci.edu/ml/datasets/Bag+of+Words

The sampler draws samples from conditional distributions over the full state space vector to infer the latent parameters indicating document topic and topic specific word occurence probabilities.