Quantifying classifier performance relative to the uninformed model
Two novel, simple metrics for classifier performance are proposed: An accuracy-based R2 and its scaled variant, the fair accuracy. They both measure the performance relative to the uninformed classifier and are better suited for imbalanced classes tha
Authors
- Igor Fischer (ORCID: https://orcid.org/0000-0001-6347-6346)
Institutions
- Heinrich Heine University Düsseldorf (DE)
Publication Details
- Published
- 2025-08-26
- DOI
- https://doi.org/10.36227/techrxiv.175624538.85171711/v1
- Primary Topic
- Machine Learning and Data Classification
- Type
- article
- Field-Weighted Citation Impact
- 0.00