A Neural Network Boosted Double Over-Dispersed Poisson Claims Reserving Model

30 Pages Posted: 22 Apr 2019

See all articles by Andrea Gabrielli

Andrea Gabrielli

ETH Zurich, Department of Mathematics, RiskLab, Students

Date Written: April 4, 2019

Abstract

We present an actuarial loss reserving technique that takes into account both claim counts and claim amounts. Separate (over-dispersed) Poisson models for the claim counts and the claim amounts are combined by a joint embedding into a neural network architecture. As starting point of the neural network calibration we use exactly these two separate (over-dispersed) Poisson models. Such a nested model can be interpreted as a boosting machine. It allows us for joint modeling and mutual learning of claim counts and claim amounts beyond the two individual (over-dispersed) Poisson models. Moreover, this choice of neural network initialization guarantees stability and accelerates representation learning.

Keywords: cross-classified over-dispersed Poisson model, claims reserving in insurance, chain-ladder reserves, double chain-ladder, neural network, embedding, claim counts, claim amounts, learning across portfolios, boosting

JEL Classification: G22, C02, C13, C15, C45, C50, C51, C52, C53

Suggested Citation

Gabrielli, Andrea, A Neural Network Boosted Double Over-Dispersed Poisson Claims Reserving Model (April 4, 2019). Available at SSRN: https://ssrn.com/abstract=3365517 or http://dx.doi.org/10.2139/ssrn.3365517

Andrea Gabrielli (Contact Author)

ETH Zurich, Department of Mathematics, RiskLab, Students ( email )

Switzerland

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
244
Abstract Views
1,316
Rank
188,987
PlumX Metrics