A systematic survey of data balancing methods that organizes the field into six taxonomic groups and reports a two-dataset case study showing method performance depends on the dataset and classifier.
Self-supervised learning is more robust to dataset imbalance.arXiv 2021
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ML 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Data Balancing Strategies: A Systematic Survey of Resampling and Augmentation Methods
A systematic survey of data balancing methods that organizes the field into six taxonomic groups and reports a two-dataset case study showing method performance depends on the dataset and classifier.