Conformal Prediction Credibility Intervals
Author:
- 2 October 2022
- journal article
- research article
- Published by Taylor & Francis in North American Actuarial Journal
- Vol. 27 (4) , 675-688
Abstract
In the predictive modeling context, the credibility estimator is a point predictor; it is easy to calculate and avoids the model misspecification risk asymptotically, but it provides no quantification of inferential uncertainty. A Bayesian prediction interval quantifies uncertainty of prediction, but it often requires expensive computation and is subject to model misspecification risk even asymptotically. Is there a way to get the best of both worlds? Based on a powerful machine learning strategy called conformal prediction, this article proposes a method that converts the credibility estimator into a conformal prediction credibility interval. This conformal prediction credibility interval contains the credibility estimator, has computational simplicity, and guarantees finite-sample validity at a pre-assigned coverage level.Keywords
Funding Information
- Casualty Actuarial Society
- Society of Actuaries
This publication has 26 references indexed in Scilit:
- Modeling Severity and Measuring Tail Risk of Norwegian Fire ClaimsNorth American Actuarial Journal, 2015
- Bayesian nonparametric predictive modeling of group health claimsInsurance: Mathematics and Economics, 2015
- A significance test for the lassoThe Annals of Statistics, 2014
- Distribution-free Prediction Bands for Non-parametric RegressionJournal of the Royal Statistical Society Series B: Statistical Methodology, 2013
- Distribution-Free Prediction SetsJournal of the American Statistical Association, 2013
- The Bernstein-Von-Mises theorem under misspecificationElectronic Journal of Statistics, 2012
- Misspecification in infinite-dimensional Bayesian statisticsThe Annals of Statistics, 2006
- Actuarial Modeling with MCMC and BUGsNorth American Actuarial Journal, 2001
- One-Sided Inference about Functionals of a DensityThe Annals of Statistics, 1988
- Experience Rating and CredibilityASTIN Bulletin, 1967