Time-Series Forecasting of Mortality Rates using Deep Learning

26 Pages Posted: 4 Jun 2020

See all articles by Francesca Perla

Francesca Perla

University of Naples Parthenope - Department of Management Studies and Quantitative Methods

Ronald Richman

QED Actuaries and Consultants

Salvatore Scognamiglio

University of Naples "Parthenope"

Mario V. Wuthrich

RiskLab, ETH Zurich

Date Written: May 6, 2020

Abstract

The time-series nature of mortality rates lends itself to processing through neural networks that are specialized to deal with sequential data, such as recurrent and convolutional networks. Although appealing intuitively, a naive implementation of these networks does not lead to enhanced predictive performance. We show how the structure of the Lee Carter model can be generalized, and propose a relatively simple convolutional network model that can be interpreted as a generalization of the Lee Carter model, allowing for its components to be evaluated in familiar terms. The model produces highly accurate forecasts on the Human Mortality Database, and, without further modification, generalizes well to the United States Mortality Database.

Keywords: Mortality Forecasting, Recurrent Neural Networks, Convolutional Neural Networks, Representation Learning, Time-Series Forecasting, Lee Carter Model, Human Mortality Database

JEL Classification: C32, C35, G22

Suggested Citation

Perla, Francesca and Richman, Ronald and Scognamiglio, Salvatore and Wuthrich, Mario V., Time-Series Forecasting of Mortality Rates using Deep Learning (May 6, 2020). Available at SSRN: https://ssrn.com/abstract=3595426 or http://dx.doi.org/10.2139/ssrn.3595426

Francesca Perla

University of Naples Parthenope - Department of Management Studies and Quantitative Methods ( email )

Via Medina 40
Via Generale Parisi, 13
Naples, 80133
United States

Ronald Richman (Contact Author)

QED Actuaries and Consultants ( email )

38 Wierda Road West
Sandton
Johannesburg, Gauteng 2196
South Africa

Salvatore Scognamiglio

University of Naples "Parthenope" ( email )

Department of Management and Quantitative Sciences
Italy

Mario V. Wuthrich

RiskLab, ETH Zurich ( email )

Department of Mathematics
Ramistrasse 101
Zurich, 8092
Switzerland

Here is the Coronavirus
related research on SSRN

Paper statistics

Downloads
170
Abstract Views
472
rank
190,651
PlumX Metrics