DPTree and DPForest: tree-based methods fulfilling demographic parity
(2026) Annals of Actuarial Science — Vol. 20, n° 1, p. 96-114 (2026)
Files
ISBA_RP_2026-09.pdf
Open Access - Adobe PDF
- 1.28 MB
Details
- Authors
- Abstract
- Tree-based methods are widely used in insurance pricing due to their simple and accurate splitting rules. However, there is no guarantee that the resulting premiums avoid indirect discrimination when features recorded in the database are correlated with the protected variable under consideration. This paper shows that splitting rules in regression trees and random forests can be adapted in order to avoid indirect discrimination related to a binary protected variable like gender. The new procedure is illustrated on motor third-party liability insurance claim data.
- Affiliations
Citations
Simon, P.-A., Denuit, M., & Trufin, J. (2026). DPTree and DPForest: tree-based methods fulfilling demographic parity. Annals of Actuarial Science, 20(1), 96-114. https://doi.org/10.1017/S1748499525100092 (Original work published 2026)