Concatenating multiple feature extraction methods

In many real-world examples, there are many ways to extract features from a dataset. Often it is beneficial to combine several methods to obtain good performance. This example shows how to use FeatureUnion to combine features obtained by PCA and univariate selection.

Combining features using this transformer has the benefit that it allows cross validation and grid searches over the whole process.

The combination used in this example is not particularly helpful on this dataset and is only used to illustrate the usage of FeatureUnion.

Out:

Combined space has 3 features
Fitting 5 folds for each of 18 candidates, totalling 90 fits
[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.
[CV] features__pca__n_components=1, features__univ_select__k=1, svm__C=0.1
[CV]  features__pca__n_components=1, features__univ_select__k=1, svm__C=0.1, score=0.933, total=   0.0s
[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:    0.0s remaining:    0.0s
[CV] features__pca__n_components=1, features__univ_select__k=1, svm__C=0.1
[CV]  features__pca__n_components=1, features__univ_select__k=1, svm__C=0.1, score=0.933, total=   0.0s
[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:    0.0s remaining:    0.0s
[CV] features__pca__n_components=1, features__univ_select__k=1, svm__C=0.1
[CV]  features__pca__n_components=1, features__univ_select__k=1, svm__C=0.1, score=0.867, total=   0.0s
[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:    0.0s remaining:    0.0s
[CV] features__pca__n_components=1, features__univ_select__k=1, svm__C=0.1
[CV]  features__pca__n_components=1, features__univ_select__k=1, svm__C=0.1, score=0.933, total=   0.0s
[Parallel(n_jobs=1)]: Done   4 out of   4 | elapsed:    0.0s remaining:    0.0s
[CV] features__pca__n_components=1, features__univ_select__k=1, svm__C=0.1
[CV]  features__pca__n_components=1, features__univ_select__k=1, svm__C=0.1, score=1.000, total=   0.0s
[Parallel(n_jobs=1)]: Done   5 out of   5 | elapsed:    0.0s remaining:    0.0s
[CV] features__pca__n_components=1, features__univ_select__k=1, svm__C=1
[CV]  features__pca__n_components=1, features__univ_select__k=1, svm__C=1, score=0.900, total=   0.0s
[Parallel(n_jobs=1)]: Done   6 out of   6 | elapsed:    0.0s remaining:    0.0s
[CV] features__pca__n_components=1, features__univ_select__k=1, svm__C=1
[CV]  features__pca__n_components=1, features__univ_select__k=1, svm__C=1, score=1.000, total=   0.0s
[Parallel(n_jobs=1)]: Done   7 out of   7 | elapsed:    0.0s remaining:    0.0s
[CV] features__pca__n_components=1, features__univ_select__k=1, svm__C=1
[CV]  features__pca__n_components=1, features__univ_select__k=1, svm__C=1, score=0.867, total=   0.0s
[Parallel(n_jobs=1)]: Done   8 out of   8 | elapsed:    0.0s remaining:    0.0s
[CV] features__pca__n_components=1, features__univ_select__k=1, svm__C=1
[CV]  features__pca__n_components=1, features__univ_select__k=1, svm__C=1, score=0.933, total=   0.0s
[Parallel(n_jobs=1)]: Done   9 out of   9 | elapsed:    0.0s remaining:    0.0s
[CV] features__pca__n_components=1, features__univ_select__k=1, svm__C=1
[CV]  features__pca__n_components=1, features__univ_select__k=1, svm__C=1, score=1.000, total=   0.0s
[CV] features__pca__n_components=1, features__univ_select__k=1, svm__C=10
[CV]  features__pca__n_components=1, features__univ_select__k=1, svm__C=10, score=0.933, total=   0.0s
[CV] features__pca__n_components=1, features__univ_select__k=1, svm__C=10
[CV]  features__pca__n_components=1, features__univ_select__k=1, svm__C=10, score=1.000, total=   0.0s
[CV] features__pca__n_components=1, features__univ_select__k=1, svm__C=10
[CV]  features__pca__n_components=1, features__univ_select__k=1, svm__C=10, score=0.900, total=   0.0s
[CV] features__pca__n_components=1, features__univ_select__k=1, svm__C=10
[CV]  features__pca__n_components=1, features__univ_select__k=1, svm__C=10, score=0.933, total=   0.0s
[CV] features__pca__n_components=1, features__univ_select__k=1, svm__C=10
[CV]  features__pca__n_components=1, features__univ_select__k=1, svm__C=10, score=1.000, total=   0.0s
[CV] features__pca__n_components=1, features__univ_select__k=2, svm__C=0.1
[CV]  features__pca__n_components=1, features__univ_select__k=2, svm__C=0.1, score=0.933, total=   0.0s
[CV] features__pca__n_components=1, features__univ_select__k=2, svm__C=0.1
[CV]  features__pca__n_components=1, features__univ_select__k=2, svm__C=0.1, score=0.967, total=   0.0s
[CV] features__pca__n_components=1, features__univ_select__k=2, svm__C=0.1
[CV]  features__pca__n_components=1, features__univ_select__k=2, svm__C=0.1, score=0.933, total=   0.0s
[CV] features__pca__n_components=1, features__univ_select__k=2, svm__C=0.1
[CV]  features__pca__n_components=1, features__univ_select__k=2, svm__C=0.1, score=0.933, total=   0.0s
[CV] features__pca__n_components=1, features__univ_select__k=2, svm__C=0.1
[CV]  features__pca__n_components=1, features__univ_select__k=2, svm__C=0.1, score=1.000, total=   0.0s
[CV] features__pca__n_components=1, features__univ_select__k=2, svm__C=1
[CV]  features__pca__n_components=1, features__univ_select__k=2, svm__C=1, score=0.933, total=   0.0s
[CV] features__pca__n_components=1, features__univ_select__k=2, svm__C=1
[CV]  features__pca__n_components=1, features__univ_select__k=2, svm__C=1, score=0.967, total=   0.0s
[CV] features__pca__n_components=1, features__univ_select__k=2, svm__C=1
[CV]  features__pca__n_components=1, features__univ_select__k=2, svm__C=1, score=0.933, total=   0.0s
[CV] features__pca__n_components=1, features__univ_select__k=2, svm__C=1
[CV]  features__pca__n_components=1, features__univ_select__k=2, svm__C=1, score=0.933, total=   0.0s
[CV] features__pca__n_components=1, features__univ_select__k=2, svm__C=1
[CV]  features__pca__n_components=1, features__univ_select__k=2, svm__C=1, score=1.000, total=   0.0s
[CV] features__pca__n_components=1, features__univ_select__k=2, svm__C=10
[CV]  features__pca__n_components=1, features__univ_select__k=2, svm__C=10, score=0.967, total=   0.0s
[CV] features__pca__n_components=1, features__univ_select__k=2, svm__C=10
[CV]  features__pca__n_components=1, features__univ_select__k=2, svm__C=10, score=0.967, total=   0.0s
[CV] features__pca__n_components=1, features__univ_select__k=2, svm__C=10
[CV]  features__pca__n_components=1, features__univ_select__k=2, svm__C=10, score=0.933, total=   0.0s
[CV] features__pca__n_components=1, features__univ_select__k=2, svm__C=10
[CV]  features__pca__n_components=1, features__univ_select__k=2, svm__C=10, score=0.933, total=   0.0s
[CV] features__pca__n_components=1, features__univ_select__k=2, svm__C=10
[CV]  features__pca__n_components=1, features__univ_select__k=2, svm__C=10, score=1.000, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=1, svm__C=0.1
[CV]  features__pca__n_components=2, features__univ_select__k=1, svm__C=0.1, score=0.933, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=1, svm__C=0.1
[CV]  features__pca__n_components=2, features__univ_select__k=1, svm__C=0.1, score=1.000, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=1, svm__C=0.1
[CV]  features__pca__n_components=2, features__univ_select__k=1, svm__C=0.1, score=0.867, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=1, svm__C=0.1
[CV]  features__pca__n_components=2, features__univ_select__k=1, svm__C=0.1, score=0.933, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=1, svm__C=0.1
[CV]  features__pca__n_components=2, features__univ_select__k=1, svm__C=0.1, score=1.000, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=1, svm__C=1
[CV]  features__pca__n_components=2, features__univ_select__k=1, svm__C=1, score=0.967, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=1, svm__C=1
[CV]  features__pca__n_components=2, features__univ_select__k=1, svm__C=1, score=1.000, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=1, svm__C=1
[CV]  features__pca__n_components=2, features__univ_select__k=1, svm__C=1, score=0.933, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=1, svm__C=1
[CV]  features__pca__n_components=2, features__univ_select__k=1, svm__C=1, score=0.933, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=1, svm__C=1
[CV]  features__pca__n_components=2, features__univ_select__k=1, svm__C=1, score=1.000, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=1, svm__C=10
[CV]  features__pca__n_components=2, features__univ_select__k=1, svm__C=10, score=0.967, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=1, svm__C=10
[CV]  features__pca__n_components=2, features__univ_select__k=1, svm__C=10, score=0.967, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=1, svm__C=10
[CV]  features__pca__n_components=2, features__univ_select__k=1, svm__C=10, score=0.900, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=1, svm__C=10
[CV]  features__pca__n_components=2, features__univ_select__k=1, svm__C=10, score=0.933, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=1, svm__C=10
[CV]  features__pca__n_components=2, features__univ_select__k=1, svm__C=10, score=1.000, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=2, svm__C=0.1
[CV]  features__pca__n_components=2, features__univ_select__k=2, svm__C=0.1, score=0.967, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=2, svm__C=0.1
[CV]  features__pca__n_components=2, features__univ_select__k=2, svm__C=0.1, score=1.000, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=2, svm__C=0.1
[CV]  features__pca__n_components=2, features__univ_select__k=2, svm__C=0.1, score=0.933, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=2, svm__C=0.1
[CV]  features__pca__n_components=2, features__univ_select__k=2, svm__C=0.1, score=0.933, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=2, svm__C=0.1
[CV]  features__pca__n_components=2, features__univ_select__k=2, svm__C=0.1, score=1.000, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=2, svm__C=1
[CV]  features__pca__n_components=2, features__univ_select__k=2, svm__C=1, score=0.967, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=2, svm__C=1
[CV]  features__pca__n_components=2, features__univ_select__k=2, svm__C=1, score=1.000, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=2, svm__C=1
[CV]  features__pca__n_components=2, features__univ_select__k=2, svm__C=1, score=0.933, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=2, svm__C=1
[CV]  features__pca__n_components=2, features__univ_select__k=2, svm__C=1, score=0.967, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=2, svm__C=1
[CV]  features__pca__n_components=2, features__univ_select__k=2, svm__C=1, score=1.000, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=2, svm__C=10
[CV]  features__pca__n_components=2, features__univ_select__k=2, svm__C=10, score=0.967, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=2, svm__C=10
[CV]  features__pca__n_components=2, features__univ_select__k=2, svm__C=10, score=1.000, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=2, svm__C=10
[CV]  features__pca__n_components=2, features__univ_select__k=2, svm__C=10, score=0.900, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=2, svm__C=10
[CV]  features__pca__n_components=2, features__univ_select__k=2, svm__C=10, score=0.933, total=   0.0s
[CV] features__pca__n_components=2, features__univ_select__k=2, svm__C=10
[CV]  features__pca__n_components=2, features__univ_select__k=2, svm__C=10, score=1.000, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=1, svm__C=0.1
[CV]  features__pca__n_components=3, features__univ_select__k=1, svm__C=0.1, score=0.967, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=1, svm__C=0.1
[CV]  features__pca__n_components=3, features__univ_select__k=1, svm__C=0.1, score=1.000, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=1, svm__C=0.1
[CV]  features__pca__n_components=3, features__univ_select__k=1, svm__C=0.1, score=0.933, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=1, svm__C=0.1
[CV]  features__pca__n_components=3, features__univ_select__k=1, svm__C=0.1, score=0.967, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=1, svm__C=0.1
[CV]  features__pca__n_components=3, features__univ_select__k=1, svm__C=0.1, score=1.000, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=1, svm__C=1
[CV]  features__pca__n_components=3, features__univ_select__k=1, svm__C=1, score=0.967, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=1, svm__C=1
[CV]  features__pca__n_components=3, features__univ_select__k=1, svm__C=1, score=1.000, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=1, svm__C=1
[CV]  features__pca__n_components=3, features__univ_select__k=1, svm__C=1, score=0.933, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=1, svm__C=1
[CV]  features__pca__n_components=3, features__univ_select__k=1, svm__C=1, score=0.967, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=1, svm__C=1
[CV]  features__pca__n_components=3, features__univ_select__k=1, svm__C=1, score=1.000, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=1, svm__C=10
[CV]  features__pca__n_components=3, features__univ_select__k=1, svm__C=10, score=1.000, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=1, svm__C=10
[CV]  features__pca__n_components=3, features__univ_select__k=1, svm__C=10, score=1.000, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=1, svm__C=10
[CV]  features__pca__n_components=3, features__univ_select__k=1, svm__C=10, score=0.933, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=1, svm__C=10
[CV]  features__pca__n_components=3, features__univ_select__k=1, svm__C=10, score=0.967, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=1, svm__C=10
[CV]  features__pca__n_components=3, features__univ_select__k=1, svm__C=10, score=1.000, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=2, svm__C=0.1
[CV]  features__pca__n_components=3, features__univ_select__k=2, svm__C=0.1, score=0.967, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=2, svm__C=0.1
[CV]  features__pca__n_components=3, features__univ_select__k=2, svm__C=0.1, score=1.000, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=2, svm__C=0.1
[CV]  features__pca__n_components=3, features__univ_select__k=2, svm__C=0.1, score=0.933, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=2, svm__C=0.1
[CV]  features__pca__n_components=3, features__univ_select__k=2, svm__C=0.1, score=0.967, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=2, svm__C=0.1
[CV]  features__pca__n_components=3, features__univ_select__k=2, svm__C=0.1, score=1.000, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=2, svm__C=1
[CV]  features__pca__n_components=3, features__univ_select__k=2, svm__C=1, score=0.967, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=2, svm__C=1
[CV]  features__pca__n_components=3, features__univ_select__k=2, svm__C=1, score=1.000, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=2, svm__C=1
[CV]  features__pca__n_components=3, features__univ_select__k=2, svm__C=1, score=0.967, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=2, svm__C=1
[CV]  features__pca__n_components=3, features__univ_select__k=2, svm__C=1, score=0.967, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=2, svm__C=1
[CV]  features__pca__n_components=3, features__univ_select__k=2, svm__C=1, score=1.000, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=2, svm__C=10
[CV]  features__pca__n_components=3, features__univ_select__k=2, svm__C=10, score=1.000, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=2, svm__C=10
[CV]  features__pca__n_components=3, features__univ_select__k=2, svm__C=10, score=1.000, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=2, svm__C=10
[CV]  features__pca__n_components=3, features__univ_select__k=2, svm__C=10, score=0.900, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=2, svm__C=10
[CV]  features__pca__n_components=3, features__univ_select__k=2, svm__C=10, score=0.967, total=   0.0s
[CV] features__pca__n_components=3, features__univ_select__k=2, svm__C=10
[CV]  features__pca__n_components=3, features__univ_select__k=2, svm__C=10, score=1.000, total=   0.0s
[Parallel(n_jobs=1)]: Done  90 out of  90 | elapsed:    0.3s finished
Pipeline(steps=[('features',
                 FeatureUnion(transformer_list=[('pca', PCA(n_components=3)),
                                                ('univ_select',
                                                 SelectKBest(k=1))])),
                ('svm', SVC(C=10, kernel='linear'))])

# Author: Andreas Mueller <amueller@ais.uni-bonn.de>
#
# License: BSD 3 clause

from sklearn.pipeline import Pipeline, FeatureUnion
from sklearn.model_selection import GridSearchCV
from sklearn.svm import SVC
from sklearn.datasets import load_iris
from sklearn.decomposition import PCA
from sklearn.feature_selection import SelectKBest

iris = load_iris()

X, y = iris.data, iris.target

# This dataset is way too high-dimensional. Better do PCA:
pca = PCA(n_components=2)

# Maybe some original features where good, too?
selection = SelectKBest(k=1)

# Build estimator from PCA and Univariate selection:

combined_features = FeatureUnion([("pca", pca), ("univ_select", selection)])

# Use combined features to transform dataset:
X_features = combined_features.fit(X, y).transform(X)
print("Combined space has", X_features.shape[1], "features")

svm = SVC(kernel="linear")

# Do grid search over k, n_components and C:

pipeline = Pipeline([("features", combined_features), ("svm", svm)])

param_grid = dict(features__pca__n_components=[1, 2, 3],
                  features__univ_select__k=[1, 2],
                  svm__C=[0.1, 1, 10])

grid_search = GridSearchCV(pipeline, param_grid=param_grid, verbose=10)
grid_search.fit(X, y)
print(grid_search.best_estimator_)

Total running time of the script: ( 0 minutes 0.301 seconds)

Gallery generated by Sphinx-Gallery