Combining feature extraction classes in scikit-learn

Question

I'm using sklearn.pipeline.Pipeline to chain feature extractors and a classifier. Is there a way to combine multiple feature selection classes (for example the ones from sklearn.feature_selection.text) in parallel and join their output?

My code right now looks as follows:

pipeline = Pipeline([
    ('vect', CountVectorizer()),
    ('tfidf', TfidfTransformer()),
    ('clf', SGDClassifier())])

It results in the following:

vect -> tfidf -> clf

I want to be able to specify a pipeline that looks as follows:

vect1 -> tfidf1 \
                 -> clf
vect2 -> tfidf2 /

score 17 · Accepted Answer · answered Oct 04 '12 at 09:36

17

This has been implemented recently in the master branch of scikit-learn under the name FeatureUnion:

http://scikit-learn.org/dev/modules/pipeline.html#feature-union

answered Oct 04 '12 at 09:36

ogrisel

39,309
12
116
125

2

Yup, I saw it at http://blog.kaggle.com/2012/09/26/impermium-andreas-blog/ after I asked the question – Daniel Oct 04 '12 at 19:04
Cool, a first user :) Let me know if you find it useful! – Andreas Mueller Oct 05 '12 at 10:29

Combining feature extraction classes in scikit-learn

1 Answers1

Linked