Re-evaluation corrects prior migraine classification baselines to macro-F1 0.71; class-dependent hybrid augmentation plus subtype aggregation reaches 0.914 with FT-Transformer, but aggregation drives most gains while the framework mainly improves average robustness across eight classifiers.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Class-Dependent Hybrid Data Augmentation for Multiclass Migraine Classification under Severe Class Imbalance
Re-evaluation corrects prior migraine classification baselines to macro-F1 0.71; class-dependent hybrid augmentation plus subtype aggregation reaches 0.914 with FT-Transformer, but aggregation drives most gains while the framework mainly improves average robustness across eight classifiers.