Join us in our upcoming webinar on: Premium control with reinforcement learning
24 AUGUST 2023 08:00 AM (EST)14:00PM (CEST)
ABSTRACT
In simplified settings, optimal premium rules can be derived by classical approaches such as dynamic programming methods providing solutions to a Bellman equation. Realistic insurance settings involve features such as reporting/payment delays and fluctuations in the number of policyholders, partly in response to varying premium levels. In such settings, classical approaches are not applicable due to the size of the state space and lack of explicit expressions for transition probabilities.
In this talk we will discuss how to design efficient algorithms in search of optimal premium rules in realistic insurance settings requiring reinforcement learning combined with function approximation. We illustrate the appropriateness of the approximate optimal premium rule compared with the true optimal premium rule in a simplified setting and further demonstrate that the approximate optimal premium rule outperforms benchmark rules in more realistic insurance settings where classical approaches fail.
SPEAKERS
Speaker Lina Palmborg
Lina Palmborg is a PhD student in insurance mathematics at Stockholm University. Her research is centered on applying and developing machine learning methods to problems in insurance mathematics, with a particular focus on reinforcement learning methods, as well as theoretical work motivated by these applications. Prior to starting her PhD, Lina worked for nearly a decade at a Swedish life insurance company, and she is a certified actuary.
Speaker Brian Fannin
Brian Fannin is a research actuary at CAS. He is also the founder of PirateGrunt LLC, a boutique consulting firm based in Durham, North Carolina specializing in predictive modeling in the property casualty markets. Fannin has been an Associate of the CAS since 2002 and a Certified Specialist in Predictive Analytics (CSPA) through The CAS Institute since 2017
ASTIN WEBINAR - The use of autoencoders for training neural networks with mixed categorical and numerical features Speaker: Łukasz Delong Session Moderator: Brian Fannin 21 SEP 2023 | 10 AM EST
