What is n_components
in sklearn.NMF
?
nmf = NMF(n_components=2, init='random', random_state=0)
nmf.fit(V)
Here's one example of how I've seen this parameter used:
When used for topic modeling in natural language processing, n_components
can be used as the number of topics to print out.
For example, if I fit the NMF object on a TF-IDF vectorizer and set n_components
= 10, I can use it to show me 10 different collections of words that I can interpret as my topics.