The LDA stands for Latent Dirichlet Allocation which is the basis for a probabilistic model of words in documents. The wikipedia article is excellent and goes

The LDA stands for Latent Dirichlet Allocation which is the basis for a probabilistic model of words in documents. The wikipedia article is excellent and goes on to explain why you would need to use Gibbs Sampling: https://en.wikipedia.org/wiki/Latent_Dirichlet_allocation https://en.wikipedia.org/wiki/Latent_Dirichlet_allocation