A new strategy for finite-sample valid prediction of future insurance claims in the regression setting

2026-01-29

AI Summary

This study addresses a critical gap in actuarial science: the absence of finite-sample valid prediction intervals for future insurance claims under regression settings. To this end, the authors propose a general transformation framework that, for the first time, extends conformal prediction methods—originally developed for unsupervised i.i.d. environments—to supervised regression scenarios. By integrating conformal inference with regression modeling, the framework guarantees finite-sample validity and enables the construction of arbitrarily many valid prediction intervals. This approach not only provides rigorous theoretical assurance but also offers practical utility, thereby filling a longstanding void in the actuarial literature concerning finite-sample valid regression-based prediction.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Other Foundations of Reasoning under Uncertainty
• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Social Networks and Social Media: Influence propagation, information diffusion, and the prediction on networks

Abstract

The extant insurance literature demonstrates a paucity of finite-sample valid prediction intervals of future insurance claims in the regression setting. To address this challenge, this article proposes a new strategy that converts a predictive method in the unsupervised iid (independent identically distributed) setting to a predictive method in the regression setting. In particular, it enables an actuary to obtain infinitely many finite-sample valid prediction intervals in the regression setting.

Problem

• finite-sample validity
• prediction intervals
• insurance claims
• regression setting

Innovation

• finite-sample validity
• prediction intervals
• regression setting
• insurance claims prediction
• unsupervised to supervised conversion