No packages match

imbalanceKit - Resampling Methods and Tools for Imbalanced Classification

A collection of resampling methods for class imbalance in classification problems. Includes the oversampling methods SMOTE with boosting (SMOTEWB) by F. Sağlam and M. A. Cengiz (2022) <doi:10.1016/j.eswa.2022.117023>, Markov chain SMOTE (MCSMOTE) by F. Sağlam and M. A. Cengiz (2025) <doi:10.1016/j.asoc.2025.113906>, uncertainty guided oversampling (UGO) by F. Sağlam and M. A. Cengiz (2026) <doi:10.1016/j.knosys.2026.115886>, and bi-directional SMOTE (BiSMOTE) by F. Sağlam and U. Yamak (2027) <doi:10.1016/j.eswa.2026.134648>, and the data relocation approach DatRel by F. Sağlam (2025) <doi:10.1007/s10994-025-06755-8>, which can also be applied to synthetic samples from any other source, together with commonly used alternatives such as SMOTE, ADASYN, Borderline-SMOTE, Safe-Level-SMOTE, Radius-SMOTE, VS-SMOTE, geometric SMOTE, random oversampling and undersampling. The methods share the same interface: they take a feature matrix and a factor class variable and return the resampled data together with the synthetic samples. The package also provides performance metrics for imbalanced classification and data complexity measures of class imbalance, class overlap and noise.

Last updated

cpp

4.00 score 3 scripts

SMOTEWB - Imbalanced Resampling using SMOTE with Boosting (SMOTEWB)

Provides the SMOTE with boosting (SMOTEWB) algorithm, where SMOTE is the synthetic minority over-sampling technique. See F. Sağlam, M. A. Cengiz (2022) <doi:10.1016/j.eswa.2022.117023>. It is a SMOTE-based resampling technique which creates synthetic data on the links between nearest neighbors. SMOTEWB uses boosting weights to determine where to generate new samples and automatically decides the number of neighbors for each sample. It is robust to noise and outperforms most of the alternatives according to Matthew Correlation Coefficient metric. Alternative resampling methods are also available in the package. This package is superseded by the 'imbalanceKit' package, which contains all of its methods and more.

Last updated

3.48 score 1 stars 1 dependents 9 scripts 358 downloads

rcccd - Class Cover Catch Digraph Classification

Fit Class Cover Catch Digraph Classification models that can be used in machine learning. Pure and proper and random walk approaches are available. Methods are explained in Priebe et al. (2001) <doi:10.1016/S0167-7152(01)00129-8>, Priebe et al. (2003) <doi:10.1007/s00357-003-0003-7>, and Manukyan and Ceyhan (2016) <doi:10.48550/arXiv.1904.04564>.

Last updated

cpp

3.18 score 1 stars 1 dependents 1 scripts 309 downloads

imbalanceDatRel - Relocated Data Oversampling for Imbalanced Data Classification

Relocates oversampled data from a specific oversampling method to cover area determined by pure and proper class cover catch digraphs (PCCCD). It prevents any data to be generated in class overlapping area. For more details, see the corresponding publication: F. Sağlam (2025) <doi:10.1007/s10994-025-06755-8>. This package is superseded by the 'imbalanceKit' package, which contains DatRel and many more resampling methods.

Last updated

2.70 score 422 downloads