Dirichlet distribution — where does it show up in ML?
hardAnswer
- : distribution over probability vectors on the (K-1)-simplex (positive components summing to 1).
- Multivariate generalization of Beta.
- Conjugate prior for the categorical / multinomial distributions.
- Uses: topic models (LDA — Latent Dirichlet Allocation), mixture-model weights, softmax priors, RL policy priors.
- Concentration parameter controls how peaked (large α) vs uniform (small α) the samples are.
Check yourself — multiple choice
- Univariate on R
- Prior on probability simplex; conjugate for categorical/multinomial; foundation of LDA, mixture weights, softmax priors
- Only for regression
- Same as Beta identically
Dirichlet: prior over prob simplex; conjugate for multinomial; used in LDA.
#distributions#bayesian
Practise Statistics Fundamentals
215 interview questions in this topic.