Python – The easiest way for getting feature names after running SelectKBest in Scikit Learn

feature-selectionpandaspythonscikit-learn

I would like to make supervised learning.

Until now I know to do supervised learning to all features.

However, I would like also to conduct experiment with the K best features.

I read the documentation and found the in Scikit learn there is SelectKBest method.

Unfortunately, I am not sure how to create new dataframe after finding those best features:

Let's assume I would like to conduct experiment with 5 best features:

from sklearn.feature_selection import SelectKBest, f_classif
select_k_best_classifier = SelectKBest(score_func=f_classif, k=5).fit_transform(features_dataframe, targeted_class)

Now if I would add the next line:

dataframe = pd.DataFrame(select_k_best_classifier)

I will receive a new dataframe without feature names (only index starting from 0 to 4).

I should replace it to:

dataframe = pd.DataFrame(fit_transofrmed_features, columns=features_names)

My question is how to create the features_names list??

I know that I should use:

 select_k_best_classifier.get_support()

Which returns array of boolean values.

The true value in the array represent the index in the right column.

How should I use this boolean array with the array of all features names I can get via the method:

feature_names = list(features_dataframe.columns.values)

Best Answer

This doesn't require loops.

# Create and fit selector
selector = SelectKBest(f_classif, k=5)
selector.fit(features_df, target)
# Get columns to keep and create new dataframe with those only
cols = selector.get_support(indices=True)
features_df_new = features_df.iloc[:,cols]

Related Solutions

Python – What’s the canonical way to check for type in Python

To check if o is an instance of str or any subclass of str, use isinstance (this would be the "canonical" way):

if isinstance(o, str):

To check if the type of o is exactly str (exclude subclasses):

if type(o) is str:

The following also works, and can be useful in some cases:

if issubclass(type(o), str):

See Built-in Functions in the Python Library Reference for relevant information.

One more note: in this case, if you're using Python 2, you may actually want to use:

if isinstance(o, basestring):

because this will also catch Unicode strings (unicode is not a subclass of str; both str and unicode are subclasses of basestring). Note that basestring no longer exists in Python 3, where there's a strict separation of strings (str) and binary data (bytes).

Alternatively, isinstance accepts a tuple of classes. This will return True if o is an instance of any subclass of any of (str, unicode):

if isinstance(o, (str, unicode)):

Python – the easiest way to remove all packages installed by pip

I've found this snippet as an alternative solution. It's a more graceful removal of libraries than remaking the virtualenv:

pip freeze | xargs pip uninstall -y

In case you have packages installed via VCS, you need to exclude those lines and remove the packages manually (elevated from the comments below):

pip freeze | grep -v "^-e" | xargs pip uninstall -y

Best Answer

Related Solutions

Python – What’s the canonical way to check for type in Python

Python – the easiest way to remove all packages installed by pip

Related Topic