Open npow opened 9 years ago
What is the correct way to limit the number features? Sort of similar to the max_features in SciKit's CountVectorizer. I was looking at the API for the BayesClassifier and it only takes stemmer and smoothing as parameters.
i found that its unstable after 1000
What is the correct way to limit the number features? Sort of similar to the max_features in SciKit's CountVectorizer. I was looking at the API for the BayesClassifier and it only takes stemmer and smoothing as parameters.