Generalized linear models and neural networks in insurance

Authors

  • A. A. Novikova National Technical University of Ukraine “Igor Sikorsky Kyiv Polytechnic Institute”, Ukraine
  • O. I. Vasylyk National Technical University of Ukraine “Igor Sikorsky Kyiv Polytechnic Institute”, Ukraine https://orcid.org/0000-0002-0880-3751

DOI:

https://doi.org/10.20535/mmtu-2025.2-083

Keywords:

insurance,, pricing,, generalized linear model,, neural network,, combined actuarial neural network,, architecture,, embedding layers,, car insurance,, Poisson regression model.

Abstract

Actuarial calculations form the basis for ensuring the solvency of the insurance company. This paper is devoted to the study of recently proposed combined actuarial neural network that combines the traditional generalized linear model used in insurance pricing with a neural network. The main idea behind using a neural network for insurance pricing is to model interactions between functions that are not covered by a generalized linear model.

Downloads

Published

2025-12-30

Issue

Vol. 2025 No. 2

Section

Application of mathematics in related sciences