Statistical agnostic regression: A machine learning method to validate regression models
Simulations demonstrate that the proposed agnostic (non-parametric) test can perform an analysis of variance comparable to the classical multivariate F-test for the slope parameter, without relying on the underlying assumptions of classical methods. A power analysis on a putative regression task revealed an overinflated false positive rate in standard ML methods, whereas the SAR test exhibited excellent control. Moreover, the residuals computed using this method represent a trade-off between those obtained from ML approaches and classical OLS.
Authors
- Francisco J. Martínez-Murcia (ORCID: https://orcid.org/0000-0001-8146-7056)
- J. M. Górriz (ORCID: https://orcid.org/0000-0001-7069-1714)
- F. Segovia (ORCID: https://orcid.org/0000-0003-1940-8834)
- John Suckling (ORCID: https://orcid.org/0000-0002-5098-1527)
- C. Jiménez-Mesa (ORCID: https://orcid.org/0000-0003-2494-2951)
- Javier Ramı́rez (ORCID: https://orcid.org/0000-0002-6229-2921)
Institutions
- Universidad de Granada (ES)
- University of Cambridge (GB)
Publication Details
- Journal
- Journal of Advanced Research
- Published
- 2025-05-01
- DOI
- https://doi.org/10.1016/j.jare.2025.04.026
- Citations
- 3
- Primary Topic
- Anomaly Detection Techniques and Applications
- Type
- article
- Field-Weighted Citation Impact
- 7.44