Package index • imbalanceKit

Methods of the author

Resampling methods developed by the package author. All oversamplers take a feature matrix and a factor class variable and return the resampled data together with the synthetic samples.

SMOTEWB(): SMOTE with Boosting (SMOTEWB, 2022)
MCSMOTE(): Markov Chain SMOTE (MCSMOTE, 2025)
BiSMOTE(): Bi-directional SMOTE (BiSMOTE, 2027)

Procedures built on a base method

DatRel moves synthetic samples back into their class region and can be applied to samples from any source. UGO uses a base oversampler together with a classifier.

DatRel(): Data Relocation (DatRel, 2025)
oversampleDatRel(): Oversampling followed by DatRel (2025)
UGO(): Uncertainty Guided Oversampling (UGO, 2026)

Other oversampling methods

Listed by year of publication.

ROS(): Random Oversampling (ROS)
SMOTE(): Synthetic Minority Oversampling Technique (SMOTE, 2002)
BLSMOTE(): Borderline-SMOTE (BLSMOTE, 2005)
ADASYN(): Adaptive Synthetic Sampling (ADASYN, 2008)
SLSMOTE(): Safe-Level SMOTE (SLSMOTE, 2009)
RWO(): Random Walk Oversampling (RWO, 2014)
ROSE(): Random Over-Sampling Examples (ROSE, 2014)
RSLSMOTE(): Relocating Safe-Level SMOTE (RSLSMOTE, 2016)
EBSMOTE(): Evaluation-Based SMOTE (EBSMOTE, 2019)
GSMOTE(): Geometric SMOTE (GSMOTE, 2019)
SyMProD(): Synthetic Minority Based on Probabilistic Distribution (SyMProD, 2020)
RadiusSMOTE(): Radius-SMOTE (RadiusSMOTE, 2021)
GQEO(): Quadrilateral Element Oversampling (GQEO, 2025)
LEAO(): Local Entropy-Adversarial Oversampling (LEAO, 2025)
VSSMOTE(): VS-SMOTE (VSSMOTE, 2025)

Undersampling

RUS(): Random Undersampling (RUS)

Performance evaluation

Metrics for imbalanced classification and a function that calculates several of them.

evaluate(): Evaluate a Classifier with Several Metrics
ACC() BACC() GMEAN() MCC() KAPPA() IAM() F1() PRECISION() RECALL() SPECIFICITY() P4() ROCAUC() PRAUC() Brier() PBrier() BBrier() ECE() BECE() LOGLOSS(): Performance Metrics for Imbalanced Classification

Data complexity measures

Measures of class imbalance, class overlap and noise, each as a separate function and all together through complexity().

complexity(): Calculate Several Complexity Measures
complexity_IR() complexity_C1() complexity_C2() complexity_ID() complexity_F1() complexity_F1v() complexity_F2() complexity_F3() complexity_F4() complexity_N1() complexity_N2() complexity_N3() complexity_N3_bal() complexity_N4() complexity_LSC() complexity_R() complexity_R_aug() complexity_SAFE() complexity_BORDERLINE() complexity_RARE() complexity_OUTLIER(): Complexity Measures for Imbalanced Classification Data

Data

banana: Banana data