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.
