Version 0.13
scikit-learn Docs · نسخهٔ 5e9184e53cd3 · BSD-3-Clause
============
Version 0.13
Version 0.13.1
February 23, 2013
The 0.13.1 release only fixes some bugs and does not add any new functionality.
Changelog
- Fixed a testing error caused by the function
cross_validation.train_test_splitbeing
interpreted as a test by Yaroslav Halchenko_.
- Fixed a bug in the reassignment of small clusters in the
cluster.MiniBatchKMeans
by Gael Varoquaux_.
- Fixed default value of `
gammaindecomposition.KernelPCAbyLars Buitinck`_.
- Updated joblib to `
0.7.0dbyGael Varoquaux`_.
- Fixed scaling of the deviance in
ensemble.GradientBoostingClassifierbyPeter Prettenhofer_.
- Better tie-breaking in
multiclass.OneVsOneClassifierbyAndreas Müller_.
- Other small improvements to tests and documentation.
People
List of contributors for release 0.13.1 by number of commits.
- 16
Lars Buitinck_ - 12
Andreas Müller_ - 8
Gael Varoquaux_ - 5 Robert Marchman
- 3
Peter Prettenhofer_ - 2 Hrishikesh Huilgolkar
- 1 Bastiaan van den Berg
- 1 Diego Molla
- 1
Gilles Louppe_ - 1
Mathieu Blondel_ - 1
Nelle Varoquaux_ - 1 Rafael Cunha de Almeida
- 1 Rolando Espinoza La fuente
- 1
Vlad Niculae_ - 1
Yaroslav Halchenko_
Version 0.13
January 21, 2013
New Estimator Classes
dummy.DummyClassifieranddummy.DummyRegressor, two
data-independent predictors by Mathieu Blondel_. Useful to sanity-check your estimators. See dummy_estimators in the user guide. Multioutput support added by Arnaud Joly_.
decomposition.FactorAnalysis, a transformer implementing the
classical factor analysis, by Christian Osendorfer_ and Alexandre Gramfort_. See FA in the user guide.
feature_extraction.FeatureHasher, a transformer implementing the
"hashing trick" for fast, low-memory feature extraction from string fields by Lars Buitinck_ and feature_extraction.text.HashingVectorizer for text documents by Olivier Grisel_ See feature_hashing and hashing_vectorizer for the documentation and sample usage.
pipeline.FeatureUnion, a transformer that concatenates
results of several other transformers by Andreas Müller_. See feature_union in the user guide.
random_projection.GaussianRandomProjection,
random_projection.SparseRandomProjection and the function random_projection.johnson_lindenstrauss_min_dim. The first two are transformers implementing Gaussian and sparse random projection matrix by Olivier Grisel_ and Arnaud Joly_. See random_projection in the user guide.
kernel_approximation.Nystroem, a transformer for approximating
arbitrary kernels by Andreas Müller_. See nystroem_kernel_approx in the user guide.
preprocessing.OneHotEncoder, a transformer that computes binary
encodings of categorical features by Andreas Müller_. See preprocessing_categorical_features in the user guide.
linear_model.PassiveAggressiveClassifierand
linear_model.PassiveAggressiveRegressor, predictors implementing an efficient stochastic optimization for linear models by Rob Zinkov_ and Mathieu Blondel_. See passive_aggressive in the user guide.
ensemble.RandomTreesEmbedding, a transformer for creating high-dimensional
sparse representations using ensembles of totally random trees by Andreas Müller_. See random_trees_embedding in the user guide.
manifold.SpectralEmbeddingand function
manifold.spectral_embedding, implementing the "laplacian eigenmaps" transformation for non-linear dimensionality reduction by Wei Li. See spectral_embedding in the user guide.
isotonic.IsotonicRegressionbyFabian Pedregosa_,Alexandre Gramfort_
and Nelle Varoquaux_,
Changelog
metrics.zero_one_loss(formerly `metrics.zero_one`) now has
an option for normalized output that reports the fraction of misclassifications, rather than the raw number of misclassifications. By Kyle Beauchamp.
tree.DecisionTreeClassifierand all derived ensemble models now
support sample weighting, by Noel Dawe_ and Gilles Louppe_.
- Speedup improvement when using bootstrap samples in forests of randomized
trees, by Peter Prettenhofer_ and Gilles Louppe_.
- Partial dependence plots for
gradient_boostingin
ensemble.partial_dependence.partial_dependence by Peter Prettenhofer_. See sphx_glr_auto_examples_inspection_plot_partial_dependence.py for an example.
- The table of contents on the website has now been made expandable by
Jaques Grobler_.
feature_selection.SelectPercentilenow breaks ties
deterministically instead of returning all equally ranked features.
feature_selection.SelectKBestand
feature_selection.SelectPercentile are more numerically stable since they use scores, rather than p-values, to rank results. This means that they might sometimes select different features than they did previously.
- Ridge regression and ridge classification fitting with `
sparse_cg` solver
no longer has quadratic memory complexity, by Lars Buitinck_ and Fabian Pedregosa_.
- Ridge regression and ridge classification now support a new fast solver
called `lsqr, by Mathieu Blondel`_.
- Speed up of
metrics.precision_recall_curveby Conrad Lee.
- Added support for reading/writing svmlight files with pairwise
preference attribute (qid in svmlight file format) in datasets.dump_svmlight_file and datasets.load_svmlight_file by Fabian Pedregosa_.
- Faster and more robust
metrics.confusion_matrixand
clustering_evaluation by Wei Li.
cross_validation.cross_val_scorenow works with precomputed kernels
and affinity matrices, by Andreas Müller_.
- LARS algorithm made more numerically stable with heuristics to drop
regressors too correlated as well as to stop the path when numerical noise becomes predominant, by Gael Varoquaux_.
- Faster implementation of
metrics.precision_recall_curveby
Conrad Lee.
- New kernel
metrics.chi2_kernelbyAndreas Müller_, often used
in computer vision applications.
- Fix of longstanding bug in
naive_bayes.BernoulliNBfixed by
Shaun Jackman.
- Implemented `
predict_probainmulticlass.OneVsRestClassifier`,
by Andrew Winterman.
- Improve consistency in gradient boosting: estimators
ensemble.GradientBoostingRegressor and ensemble.GradientBoostingClassifier use the estimator tree.DecisionTreeRegressor instead of the tree._tree.Tree data structure by Arnaud Joly_.
- Fixed a floating point exception in the
decision trees <tree>
module, by Seberg.
- Fix
metrics.roc_curvefails when y_true has only one class
by Wei Li.
- Add the
metrics.mean_absolute_errorfunction which computes the
mean absolute error. The metrics.mean_squared_error, metrics.mean_absolute_error and metrics.r2_score metrics support multioutput by Arnaud Joly_.
- Fixed `
class_weightsupport insvm.LinearSVC` and
linear_model.LogisticRegression by Andreas Müller_. The meaning of `class_weight` was reversed as erroneously higher weight meant less positives of a given class in earlier releases.
- Improve narrative documentation and consistency in
sklearn.metrics for regression and classification metrics by Arnaud Joly_.
- Fixed a bug in
sklearn.svm.SVCwhen using csr-matrices with
unsorted indices by Xinfan Meng and Andreas Müller_.
cluster.MiniBatchKMeans: Add random reassignment of cluster centers
with little observations attached to them, by Gael Varoquaux_.
API changes summary
- Renamed all occurrences of `
n_atomston_components` for consistency.
This applies to decomposition.DictionaryLearning, decomposition.MiniBatchDictionaryLearning, decomposition.dict_learning, decomposition.dict_learning_online.
- Renamed all occurrences of `
max_iterstomax_iter` for consistency.
This applies to semi_supervised.LabelPropagation and semi_supervised.label_propagation.LabelSpreading.
- Renamed all occurrences of `
learn_ratetolearning_rate` for
consistency in ensemble.BaseGradientBoosting and ensemble.GradientBoostingRegressor.
- The module `
sklearn.linear_model.sparse` is gone. Sparse matrix support
was already integrated into the "regular" linear models.
sklearn.metrics.mean_square_error, which incorrectly returned the
accumulated error, was removed. Use metrics.mean_squared_error instead.
- Passing `
class_weightparameters tofit` methods is no longer
supported. Pass them to estimator constructors instead.
- GMMs no longer have `
decodeandrvsmethods. Use thescore`,
`predict or sample` methods instead.
- The `
solver` fit option in Ridge regression and classification is now
deprecated and will be removed in v0.14. Use the constructor option instead.
feature_extraction.text.DictVectorizernow returns sparse
matrices in the CSR format, instead of COO.
- Renamed `
kincross_validation.KFold` and
cross_validation.StratifiedKFold to `n_folds, renamed n_bootstraps to n_iter in cross_validation.Bootstrap`.
- Renamed all occurrences of `
n_iterationston_iter` for consistency.
This applies to cross_validation.ShuffleSplit, cross_validation.StratifiedShuffleSplit, utils.extmath.randomized_range_finder and utils.extmath.randomized_svd.
- Replaced `
rhoinlinear_model.ElasticNet` and
linear_model.SGDClassifier by `l1_ratio. The rho parameter had different meanings; l1_ratio was introduced to avoid confusion. It has the same meaning as previously rho in linear_model.ElasticNet and (1-rho) in linear_model.SGDClassifier`.
linear_model.LassoLarsandlinear_model.Larsnow
store a list of paths in the case of multiple targets, rather than an array of paths.
- The attribute `
gmmofhmm.GMMHMMwas renamed togmm_`
to adhere more strictly with the API.
cluster.spectral_embeddingwas moved to
manifold.spectral_embedding.
- Renamed `
eig_tolinmanifold.spectral_embedding`,
cluster.SpectralClustering to `eigen_tol, renamed mode to eigen_solver`.
- Renamed `
modeinmanifold.spectral_embedding` and
cluster.SpectralClustering to `eigen_solver`.
- `
classes_andn_classes_` attributes of
tree.DecisionTreeClassifier and all derived ensemble models are now flat in case of single output problems and nested in case of multi-output problems.
- The `
estimators_` attribute of
ensemble.GradientBoostingRegressor and ensemble.GradientBoostingClassifier is now an array of tree.DecisionTreeRegressor.
- Renamed `
chunk_sizetobatch_size` in
decomposition.MiniBatchDictionaryLearning and decomposition.MiniBatchSparsePCA for consistency.
svm.SVCandsvm.NuSVCnow provide a `classes_`
attribute and support arbitrary dtypes for labels `y. Also, the dtype returned by predict now reflects the dtype of y during fit (used to be np.float`).
- Changed default test_size in
cross_validation.train_test_split
to None, added possibility to infer `test_size from train_size in cross_validation.ShuffleSplit and cross_validation.StratifiedShuffleSplit`.
- Renamed function
sklearn.metrics.zero_oneto
sklearn.metrics.zero_one_loss. Be aware that the default behavior in sklearn.metrics.zero_one_loss is different from sklearn.metrics.zero_one: `normalize=False is changed to normalize=True`.
- Renamed function
metrics.zero_one_scoreto
metrics.accuracy_score.
datasets.make_circlesnow has the same number of inner and outer points.
- In the Naive Bayes classifiers, the `
class_prior` parameter was moved
from `fit to __init__`.
People
List of contributors for release 0.13 by number of commits.
- 364
Andreas Müller_ - 143
Arnaud Joly_ - 137
Peter Prettenhofer_ - 131
Gael Varoquaux_ - 117
Mathieu Blondel_ - 108
Lars Buitinck_ - 106 Wei Li
- 101
Olivier Grisel_ - 65
Vlad Niculae_ - 54
Gilles Louppe_ - 40
Jaques Grobler_ - 38
Alexandre Gramfort_ - 30
Rob Zinkov_ - 19 Aymeric Masurelle
- 18 Andrew Winterman
- 17
Fabian Pedregosa_ - 17 Nelle Varoquaux
- 16
Christian Osendorfer_ - 14
Daniel Nouri_ - 13
Virgile Fritsch <VirgileFritsch> - 13 syhw
- 12
Satrajit Ghosh_ - 10 Corey Lynch
- 10 Kyle Beauchamp
- 9 Brian Cheung
- 9 Immanuel Bayer
- 9 mr.Shu
- 8 Conrad Lee
- 8
James Bergstra_ - 7 Tadej Janež
- 6 Brian Cajes
- 6
Jake Vanderplas_ - 6 Michael
- 6 Noel Dawe
- 6 Tiago Nunes
- 6 cow
- 5 Anze
- 5 Shiqiao Du
- 4 Christian Jauvin
- 4 Jacques Kvam
- 4 Richard T. Guy
- 4
Robert Layton_ - 3 Alexandre Abraham
- 3 Doug Coleman
- 3 Scott Dickerson
- 2 ApproximateIdentity
- 2 John Benediktsson
- 2 Mark Veronda
- 2 Matti Lyra
- 2 Mikhail Korobov
- 2 Xinfan Meng
- 1 Alejandro Weinstein
- 1
Alexandre Passos_ - 1 Christoph Deil
- 1 Eugene Nizhibitsky
- 1 Kenneth C. Arnold
- 1 Luis Pedro Coelho
- 1 Miroslav Batchkarov
- 1 Pavel
- 1 Sebastian Berg
- 1 Shaun Jackman
- 1 Subhodeep Moitra
- 1 bob
- 1 dengemann
- 1 emanuele
- 1 x006