1.13. Feature selection#

The classes in the sklearn.feature_selection module can be used for feature selection/dimensionality reduction on sample sets, either to improve estimators’ accuracy scores or to boost their performance on very high-dimensional datasets.

1.13.1. Removing features with low variance#

VarianceThreshold is a simple baseline approach to feature selection. It removes all features whose variance doesn’t meet some threshold. By default, it removes all zero-variance features, i.e. features that have the same value in all samples.

As an example, suppose that we have a dataset with boolean features, and we want to remove all features that are either one or zero (on or off) in more than 80% of the samples. Boolean features are Bernoulli random variables, and the variance of such variables is given by

\[\mathrm{Var}[X] = p(1 - p)\]

so we can select using the threshold .8 * (1 - .8):

>>> from sklearn.feature_selection import VarianceThreshold
>>> X = [[0, 0, 1], [0, 1, 0], [1, 0, 0], [0, 1, 1], [0, 1, 0], [0, 1, 1]]
>>> sel = VarianceThreshold(threshold=(.8 * (1 - .8)))
>>> sel.fit_transform(X)
array([[0, 1],
       [1, 0],
       [0, 0],
       [1, 1],
       [1, 0],
       [1, 1]])

As expected, VarianceThreshold has removed the first column, which has a probability \(p = 5/6 > .8\) of containing a zero.

1.13.2. Univariate feature selection#

Univariate feature selection works by selecting the best features based on univariate statistical tests. It can be seen as a preprocessing step to an estimator. Scikit-learn exposes feature selection routines as objects that implement the transform method:

  • SelectKBest removes all but the \(k\) highest scoring features

  • SelectPercentile removes all but a user-specified highest scoring percentage of features

  • using common univariate statistical tests for each feature: false positive rate SelectFpr, false discovery rate SelectFdr, or family wise error SelectFwe.

  • GenericUnivariateSelect allows to perform univariate feature selection with a configurable strategy. This allows to select the best univariate selection strategy with hyper-parameter search estimator.

For instance, we can use a F-test to retrieve the two best features for a dataset as follows:

>>> from sklearn.datasets import load_iris
>>> from sklearn.feature_selection import SelectKBest
>>> from sklearn.feature_selection import f_classif
>>> X, y = load_iris(return_X_y=True)
>>> X.shape
(150, 4)
>>> X_new = SelectKBest(f_classif, k=2).fit_transform(X, y)
>>> X_new.shape
(150, 2)

These objects take as input a scoring function that returns univariate scores and p-values (or only scores for SelectKBest and SelectPercentile):

The methods based on F-test estimate the degree of linear dependency between two random variables. On the other hand, mutual information methods can capture any kind of statistical dependency, but being nonparametric, they require more samples for accurate estimation. Note that the \(\chi^2\)-test should only be applied to non-negative features, such as frequencies.


Beware not to use a regression scoring function with a classification problem, you will get useless results.


The SelectPercentile and SelectKBest support unsupervised feature selection as well. One needs to provide a score_func where y=None. The score_func should use internally X to compute the scores.


1.13.3. Recursive feature elimination#

Given an external estimator that assigns weights to features (e.g., the coefficients of a linear model), the goal of recursive feature elimination (RFE) is to select features by recursively considering smaller and smaller sets of features. First, the estimator is trained on the initial set of features and the importance of each feature is obtained either through any specific attribute (such as coef_, feature_importances_) or callable. Then, the least important features are pruned from current set of features. That procedure is recursively repeated on the pruned set until the desired number of features to select is eventually reached.

RFECV performs RFE in a cross-validation loop to find the optimal number of features. In more details, the number of features selected is tuned automatically by fitting an RFE selector on the different cross-validation splits (provided by the cv parameter). The performance of the RFE selector are evaluated using scorer for different number of selected features and aggregated together. Finally, the scores are averaged across folds and the number of features selected is set to the number of features that maximize the cross-validation score.


1.13.4. Feature selection using SelectFromModel#

SelectFromModel is a meta-transformer that can be used alongside any estimator that assigns importance to each feature through a specific attribute (such as coef_, feature_importances_) or via an importance_getter callable after fitting. The features are considered unimportant and removed if the corresponding importance of the feature values are below the provided threshold parameter. Apart from specifying the threshold numerically, there are built-in heuristics for finding a threshold using a string argument. Available heuristics are “mean”, “median” and float multiples of these like “0.1*mean”. In combination with the threshold criteria, one can use the max_features parameter to set a limit on the number of features to select.

For examples on how it is to be used refer to the sections below.

Examples L1-based feature selection#

Linear models penalized with the L1 norm have sparse solutions: many of their estimated coefficients are zero. When the goal is to reduce the dimensionality of the data to use with another classifier, they can be used along with SelectFromModel to select the non-zero coefficients. In particular, sparse estimators useful for this purpose are the Lasso for regression, and of LogisticRegression and LinearSVC for classification:

>>> from sklearn.svm import LinearSVC
>>> from sklearn.datasets import load_iris
>>> from sklearn.feature_selection import SelectFromModel
>>> X, y = load_iris(return_X_y=True)
>>> X.shape
(150, 4)
>>> lsvc = LinearSVC(C=0.01, penalty="l1", dual=False).fit(X, y)
>>> model = SelectFromModel(lsvc, prefit=True)
>>> X_new = model.transform(X)
>>> X_new.shape
(150, 3)

With SVMs and logistic-regression, the parameter C controls the sparsity: the smaller C the fewer features selected. With Lasso, the higher the alpha parameter, the fewer features selected.


L1-recovery and compressive sensing#

For a good choice of alpha, the Lasso can fully recover the exact set of non-zero variables using only few observations, provided certain specific conditions are met. In particular, the number of samples should be “sufficiently large”, or L1 models will perform at random, where “sufficiently large” depends on the number of non-zero coefficients, the logarithm of the number of features, the amount of noise, the smallest absolute value of non-zero coefficients, and the structure of the design matrix X. In addition, the design matrix must display certain specific properties, such as not being too correlated.

There is no general rule to select an alpha parameter for recovery of non-zero coefficients. It can by set by cross-validation (LassoCV or LassoLarsCV), though this may lead to under-penalized models: including a small number of non-relevant variables is not detrimental to prediction score. BIC (LassoLarsIC) tends, on the opposite, to set high values of alpha.


Richard G. Baraniuk “Compressive Sensing”, IEEE Signal Processing Magazine [120] July 2007 http://users.isr.ist.utl.pt/~aguiar/CS_notes.pdf Tree-based feature selection#

Tree-based estimators (see the sklearn.tree module and forest of trees in the sklearn.ensemble module) can be used to compute impurity-based feature importances, which in turn can be used to discard irrelevant features (when coupled with the SelectFromModel meta-transformer):

>>> from sklearn.ensemble import ExtraTreesClassifier
>>> from sklearn.datasets import load_iris
>>> from sklearn.feature_selection import SelectFromModel
>>> X, y = load_iris(return_X_y=True)
>>> X.shape
(150, 4)
>>> clf = ExtraTreesClassifier(n_estimators=50)
>>> clf = clf.fit(X, y)
>>> clf.feature_importances_  
array([ 0.04...,  0.05...,  0.4...,  0.4...])
>>> model = SelectFromModel(clf, prefit=True)
>>> X_new = model.transform(X)
>>> X_new.shape               
(150, 2)


1.13.5. Sequential Feature Selection#

Sequential Feature Selection [sfs] (SFS) is available in the SequentialFeatureSelector transformer. SFS can be either forward or backward:

Forward-SFS is a greedy procedure that iteratively finds the best new feature to add to the set of selected features. Concretely, we initially start with zero features and find the one feature that maximizes a cross-validated score when an estimator is trained on this single feature. Once that first feature is selected, we repeat the procedure by adding a new feature to the set of selected features. The procedure stops when the desired number of selected features is reached, as determined by the n_features_to_select parameter.

Backward-SFS follows the same idea but works in the opposite direction: instead of starting with no features and greedily adding features, we start with all the features and greedily remove features from the set. The direction parameter controls whether forward or backward SFS is used.

Details on Sequential Feature Selection#

In general, forward and backward selection do not yield equivalent results. Also, one may be much faster than the other depending on the requested number of selected features: if we have 10 features and ask for 7 selected features, forward selection would need to perform 7 iterations while backward selection would only need to perform 3.

SFS differs from RFE and SelectFromModel in that it does not require the underlying model to expose a coef_ or feature_importances_ attribute. It may however be slower considering that more models need to be evaluated, compared to the other approaches. For example in backward selection, the iteration going from m features to m - 1 features using k-fold cross-validation requires fitting m * k models, while RFE would require only a single fit, and SelectFromModel always just does a single fit and requires no iterations.



1.13.6. Feature selection as part of a pipeline#

Feature selection is usually used as a pre-processing step before doing the actual learning. The recommended way to do this in scikit-learn is to use a Pipeline:

clf = Pipeline([
  ('feature_selection', SelectFromModel(LinearSVC(penalty="l1"))),
  ('classification', RandomForestClassifier())
clf.fit(X, y)

In this snippet we make use of a LinearSVC coupled with SelectFromModel to evaluate feature importances and select the most relevant features. Then, a RandomForestClassifier is trained on the transformed output, i.e. using only relevant features. You can perform similar operations with the other feature selection methods and also classifiers that provide a way to evaluate feature importances of course. See the Pipeline examples for more details.