TfidfVectorizer

Import

from sklearn.feature_extraction.text import TfidfVectorizer

Vectorise corpus

corpus = [] # list of strings

vectorizer = TfidfVectorizer()

X = vectorizer.fit_transform(corpus)

View the features (aka the words it thinks are important)

# print the following
vectorizer.get_feature_names()

Remove words that are pointless

# change the code above to
vectorizer = TfidfVectorizer(max_df=0.5) 
# this will only look at words that are used infrequently