A fixed-value (asymmetric) dirichlet distribution, corresponding to add psuedocount smoothing.
A parametric topic model with a fixed number of (named) topics.
Description of the Dirichlet hyperparmeters for a multinomial distribution.
Used add-k asymmetric dirichlet term smoothing.
Used add-k asymmetric dirichlet topic smoothing.
A document with a single hard assignment (Int) for each observation in a set of observations.
Abstract base class of a Gibbs sampler for LDA.
Hard assignments-based model keeps int counts; base trait of Gibbs sampler based topic models.
Counts as hard assignment model state.
Document with per-term distribution over topics.
Soft assignments-based model keeps distributions for each word observation.
Distributions as soft assignment model state.
A fixed-value (symmetric) dirichlet distribution, corresponding to add psuedocount smoothing.
Implementation trait for topic models.
ModelCompanion implementation for topic models.