Another look at the zero integral difference between lorenz and concentration curves in supervised learning

Michel Denuit () and Julien Trufin
Additional contact information
Michel Denuit: Université catholique de Louvain, LIDAM/ISBA, Belgium
Julien Trufin: ULB

No 2025026, LIDAM Discussion Papers ISBA from Université catholique de Louvain, Institute of Statistics, Biostatistics and Actuarial Sciences (ISBA)

Abstract: We revisit the area between the concentration and Lorenz curves (ABC) criterion for model assessment in supervised learning. Insurance pricing is considered throughout the paper to illustrate the concepts but the results apply to any other setting where the mean of a response must be estimated from data. Building on the characterization of these curves, we provide new equivalent formulations for the case where the ABC vanishes. First, we characterize a vanishing ABC as the absence of correlation between pricing error and the ranks induced by the candidate premiums, making the link with Gini and Co-Gini coefficients. Inboth the discrete and continuous cases, we then show that a vanishing ABC corresponds toglobal balance in a modified portfolio that overweights lower premium classes. These results complement existing work on auto-calibration and contribute to a better understanding of ABC as a diagnostic tool in insurance pricing and related applications.

Keywords: Insurance pricing; Lorenz curve; concentration curve; area between the curves (search for similar items in EconPapers)
Pages: 17
Date: 2025-12-11
New Economics Papers: this item is included in nep-com
References: Add references at CitEc
Citations:

Downloads: (external link)
https://dial.uclouvain.be/pr/boreal/en/object/bore ... tastream/PDF_01/view (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:aiz:louvad:2025026

Access Statistics for this paper

More papers in LIDAM Discussion Papers ISBA from Université catholique de Louvain, Institute of Statistics, Biostatistics and Actuarial Sciences (ISBA) Voie du Roman Pays 20, 1348 Louvain-la-Neuve (Belgium). Contact information at EDIRC.
Bibliographic data for series maintained by Nadja Peiffer ().