Backtesting Recurrent Neural Networks with Gated Recurrent Unit: Probing with Chilean Mortality Data

Research output: Chapter in Book/Report/Conference proceedingChapterpeer-review

1 Citation (Scopus)

Abstract

Understanding the survival prospects of a given population is essential in multiple research and policy areas, including public and private health care and social care, demographic analysis, pension systems evaluation, the valuation of life insurance and retirement income contracts, and the pricing and risk management of novel longevity-linked capital market instruments. This paper conducts a backtesting analysis to assess the predictive performance of Recurrent Neural Networks (RNN) with Gated Recurrent Unit (GRU) architecture in modelling and multivariate time series forecasting of age-specific mortality rates on Chilean mortality data. We investigate the best specification for one, two, and three hidden layers GRU networks and compare the RNN’s forecasting accuracy with that produced by principal component methods, namely a Regularized Singular Value Decomposition (RSVD) model. The empirical results suggest that the forecasting accuracy of RNN models critically depends on hyperparameter calibration and that the two hidden layer RNN-GRU networks outperform the RSVD model. RNNs can generate mortality schedules that are biologically plausible and fit well the mortality schedules across age and time. However, further investigation is necessary to confirm the superiority of deep learning methods in forecasting human survival across different populations and periods.
Original languageEnglish
Title of host publicationAdvances and Applications in Computer Science, Electronics, and Industrial Engineering
Subtitle of host publicationProceedings of the Conference on Computer Science, Electronics and Industrial Engineering (CSEI 2021)
EditorsMarcelo V. Garcia, Félix Fernández-Peña, Carlos Gordón-Gallegos
PublisherSpringer
Chapter9
Pages159-174
Number of pages16
ISBN (Electronic)978-3-030-97718-4
ISBN (Print)978-3-030-97718-4
DOIs
Publication statusPublished - 26 May 2022
EventInternational Conference on Computer Science, Electronics and Industrial Engineering (CSEI) - Virtual
Duration: 25 Oct 202129 Oct 2021
Conference number: 3rd
http://csei.uta.edu.ec/csei2021/

Publication series

NameLecture Notes in Networks and Systems
Volume433
ISSN (Print)2367-3370
ISSN (Electronic)2367-3389

Conference

ConferenceInternational Conference on Computer Science, Electronics and Industrial Engineering (CSEI)
Abbreviated titleCSEI 2021
Period25/10/2129/10/21
Internet address

Keywords

  • Recurrent Neural Networks (RNN)
  • Mortality modelling and forecasting
  • Life insurance
  • Backtesting

Fingerprint

Dive into the research topics of 'Backtesting Recurrent Neural Networks with Gated Recurrent Unit: Probing with Chilean Mortality Data'. Together they form a unique fingerprint.

Cite this