Version 0.5.0#
June 28, 2019
Changelog#
Changed models#
The following models or function might give different results even if the
same data X and y are the same.
imblearn.ensemble.RUSBoostClassifierdefault estimator changed fromsklearn.tree.DecisionTreeClassifierwith full depth to a decision stump (i.e., tree withmax_depth=1).
Documentation#
Correct the definition of the ratio when using a
floatin sampling strategy for the over-sampling and under-sampling. #525 by Ariel Rossanigo.Add
imblearn.over_sampling.BorderlineSMOTEandimblearn.over_sampling.SVMSMOTEin the API documenation. #530 by Guillaume Lemaitre.
Enhancement#
Add Parallelisation for SMOTEENN and SMOTETomek. #547 by Michael Hsieh.
Add
imblearn.utils._show_versions. Updated the contribution guide and issue template showing how to print system and dependency information from the command line. #557 by Alexander L. Hayes.Add
imblearn.over_sampling.KMeansSMOTEwhich is an over-sampler clustering points before to apply SMOTE. #435 by Stephan Heijl.
Maintenance#
Make it possible to
import imblearnand access submodule. #500 by Guillaume Lemaitre.Remove support for Python 2, remove deprecation warning from scikit-learn 0.21. #576 by Guillaume Lemaitre.
Bug#
Fix wrong usage of
keras.layers.BatchNormalizationinporto_seguro_keras_under_sampling.pyexample. The batch normalization was moved before the activation function and the bias was removed from the dense layer. #531 by Guillaume Lemaitre.Fix bug which converting to COO format sparse when stacking the matrices in
imblearn.over_sampling.SMOTENC. This bug was only old scipy version. #539 by Guillaume Lemaitre.Fix bug in
imblearn.pipeline.Pipelinewhere None could be the final estimator. #554 by Oliver Rausch.Fix bug in
imblearn.over_sampling.SVMSMOTEandimblearn.over_sampling.BorderlineSMOTEwhere the default parameter ofn_neighborswas not set properly. #578 by Guillaume Lemaitre.Fix bug by changing the default depth in
imblearn.ensemble.RUSBoostClassifierto get a decision stump as a weak learner as in the original paper. #545 by Christos Aridas.Allow to import
kerasdirectly fromtensorflowin theimblearn.keras. #531 by Guillaume Lemaitre.