Under-sampling Samplers

The imbalanced_ensemble.sampler.under_sampling submodule contains methods to under-sample a dataset.

ClusterCentroids(**kwargs)

Undersample by generating centroids based on clustering methods.

RandomUnderSampler(**kwargs)

Class to perform random under-sampling.

InstanceHardnessThreshold(**kwargs)

Undersample based on the instance hardness threshold.

NearMiss(**kwargs)

Class to perform under-sampling based on NearMiss methods.

TomekLinks(**kwargs)

Under-sampling by removing Tomek’s links.

EditedNearestNeighbours(**kwargs)

Undersample based on the edited nearest neighbour method.

RepeatedEditedNearestNeighbours(**kwargs)

Undersample based on the repeated edited nearest neighbour method.

AllKNN(**kwargs)

Undersample based on the AllKNN method.

OneSidedSelection(**kwargs)

Class to perform under-sampling based on one-sided selection method.

CondensedNearestNeighbour(**kwargs)

Undersample based on the condensed nearest neighbour method.

NeighbourhoodCleaningRule(**kwargs)

Undersample based on the neighbourhood cleaning rule.

BalanceCascadeUnderSampler(**kwargs)

Class to perform under-sampling for BalanceCascade in [R4d2584dbf2bf-1].

SelfPacedUnderSampler(**kwargs)

Class to perform self-paced under-sampling in [R58fe15894bd9-1].