Benchmark Imbalanced Datasets Used for Evaluating a Fast Density-Based Undersampling Algorithm

This record provides a collection of publicly available benchmark datasets used to evaluate a fast density-based undersampling algorithm for imbalanced and overlapping data. The datasets were obtained from well-known public repositories, including the UCI Machine Learning Repository, the KDD Cup archive, and the KEEL Data-Mining Software Tool Repository. They cover a wide range of data sizes, imbalance ratios, and overlap levels, and were used to assess the efficiency, scalability, and classification performance of the proposed method across different classifiers. This collection is intended to support reproducibility and facilitate comparative studies in imbalanced learning research.