Incorporating Industry Stylized Facts into Mortality Tables: Transfer Learning with Monotonicity Constraints

38 Pages Posted: 18 Nov 2021 Last revised: 1 May 2024

See all articles by Hong Beng Lim

Hong Beng Lim

The University of Iowa, Department of Statistics and Actuarial Science

N. D. Shyamalkumar

The University of Iowa

Date Written: October 15, 2021

Abstract

Developing mortality tables can be challenging when a firm lacks credible experience data or the expertise to adjust such rates, especially for smaller life firms or underwriters serving the senior life settlements industry. The recently proposed entity embedding neural network (EENN) methodology aims to resolve this issue by extracting features from industry tables reflecting such information for use in fitting mortality rates for such firms. Notably, such fitted rates do not incorporate desired monotonicity constraints e.g. with respect to attained age. We extend the EENN methodology to do so, which has the added benefit of regularizing rates and, as we show, improving predictive performance.

Keywords: Mortality modeling, transfer learning, entity embeddings, neural network, monotonic regression, monotonic network

Suggested Citation

Lim, Hong Beng and Shyamalkumar, Nariankadu, Incorporating Industry Stylized Facts into Mortality Tables: Transfer Learning with Monotonicity Constraints (October 15, 2021). Available at SSRN: https://ssrn.com/abstract=3964181 or http://dx.doi.org/10.2139/ssrn.3964181

Hong Beng Lim (Contact Author)

The University of Iowa, Department of Statistics and Actuarial Science ( email )

Iowa City, IA 52242-1409
United States

Nariankadu Shyamalkumar

The University of Iowa ( email )

Iowa City, IA 52242-1409
United States

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

Paper statistics

Downloads
61
Abstract Views
410
Rank
663,185
PlumX Metrics