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

• Published in: Data mining and knowledge discovery 2021-11, Vol.35 (6), p.2679-2713
• Main Authors: Bernardo, Alessio, Della Valle, Emanuele
• Format: Article
• Language: English
• Subjects: Algorithms; Artificial Intelligence; Chemistry and Earth Sciences; Classification; Computer Science; Data mining; Data Mining and Knowledge Discovery; Data transmission; Drift; Information Storage and Retrieval; Machine learning; Oversampling; Physics; Pipelines; Sampling techniques; Special Issue of the Journal Track of ECML PKDD 2021; Statistics for Engineering
• ISSN: 1384-5810; 1573-756X
• DOI: 10.1007/s10618-021-00786-0

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