LDA vs NMF for topic modeling — which do you pick?
mediumAnswer
- LDA: probabilistic — documents = Dirichlet mixture of topics, topics = Dirichlet mixture over words.
- Fit via variational Bayes or Gibbs.
- Interpretable but sensitive to hyperparameters (α, β).
- NMF: matrix factorization of TF-IDF, faster and often more coherent on short text.
- LDA better for longer documents / traditional corpora; NMF better for short noisy text (tweets, reviews).
- Both superseded by embedding-based topic models (BERTopic) on modern short text.
Check yourself — multiple choice
- Same thing
- LDA: probabilistic Dirichlet mixture, good for long docs; NMF: TF-IDF matrix factorization, faster + often more coherent on short text; BERTopic modern default
- Random
- Only LDA
LDA: long docs; NMF: short/noisy text; BERTopic dominates now.
#nlp#text#clustering
Practise Unsupervised Learning
214 interview questions in this topic.