VFC-SMOTE: very fast continuous synthetic minority oversampling for evolving data streams

Abstract The world is constantly changing, and so are the massive amount of data produced. However, only a few studies deal with online class imbalance learning that combines challenges class-imbalanced streams concept drift. In this paper, we propose very fast continuous synthetic minority oversampling technique ( VFC - SMOTE ). It novel meta-strategy to be prepended any streaming machine classification algorithm aiming at using new version Smote Borderline inspired by Data Sketching. We benchmarked pipelines on real containing different drifts, levels, distributions. bring statistical evidence learn models whose performances better than state-of-the-art. Moreover, analyze time/memory consumption drift recovery speed.