ABSTRACT
This paper proposes an empirical validation framework for analyzing the convergence rates of Kullback-Leibler barycentre algorithms on real-world actuarial datasets, addressing heavy-tailed distributions and heterogeneous risk profiles.
PAPER · PDF
Loading PDF...
Key findings
Develops a research methodology combining problem formulation, Iterative Bregman Projections algorithm, and comprehensive evaluation metrics.
Addresses research gaps at the intersection of optimal transport theory and actuarial science.
Provides practical guidance for implementing barycenter-based methods in insurance risk modeling.
Limitations & open questions
Theoretical results rely on assumptions that may not hold in actuarial applications.
Further research needed to address temporal dependencies in insurance data.