Answers are grounded in this paper and cite their sources.
Develops nonparametric dependence modeling in count and hybrid insurance data, suggesting enhanced accuracy for joint distributions.
Accurate modeling of dependence in insurance data is essential for pricing, reserving, and capital allocation, particularly when claim frequencies and severities interact in complex ways. Classical independence assumptions often fail in practice, and standard copula models face challenges when applied to discrete or mixed data. This paper develops a nonparametric framework for dependence modeling in count–count and hybrid frequency–severity settings using Sklar-type density estimation. Two exact Sklar-type decompositions clarify the structure of dependence in discrete data, while a kernel-based analogue and a hybrid estimator provide flexible, fully nonparametric tools for discrete and mixed insurance variables. A comprehensive simulation study across multiple sample sizes and dependence levels evaluates estimator performance using several divergence measures. An application to a motor-insurance dataset demonstrates that the proposed methods capture dependence patterns with high fidelity and yield accurate joint distribution estimates. The results highlight the practical value of nonparametric Sklar-type estimators for contemporary actuarial modeling involving discrete or hybrid data.
No takes yet. Share an insight, caveat, or question.
Ghobadlou et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: