Announcement: SOA congratulates the new ASAs and CERAs for November 2024.

Insurance Data Science: A Collection of Cutting-Edge Research in the Annals of Actuarial Science

By Andreas Tsanakas

Expanding Horizons, August 2021

Virtual touchscreen project management data

A special issue of the Annals of Actuarial Science, focusing on insurance data science, was published on July 1. As explained by the editors—Katrien Antonio (KU Leuven), Christophe Dutang (Université Paris Dauphine), and Andreas Tsanakas (City, University of London)—the theme of this issue grew out of the rapid and transformative developments in the field, with the increased use of computational statistics, machine learning and artificial intelligence models in insurance applications.[1] While the special issue has been associated with the Second Insurance Data Science Conference, held in June 2019 at ETH Zurich, the published papers come from a wider and very distinguished pool of authors.

Artificial intelligence methods, particularly deep neural networks, are being increasingly and extensively deployed in different areas of actuarial science. In a two-part article, Richman offers a wide-ranging review, spanning areas from mortality modeling to claims reserving and telematics.[2] Fernadez-Arjona shows how neural networks can be used to construct proxy models in the context of risk neutral pricing of variable annuities.[3] Zhu and Wüthrich combine unsupervised learning techniques with a pretrained convolutional neural network used for image processing to cluster driving styles.[4]

A particular focus of cutting-edge statistical methods has been the modeling of mortality. Peters et al.  study persistence and long-term memory of mortality data, linking these properties to fractional Brownian motion and multifractality.[5] Huynh and Ludkovski deploy Gaussian Processes to model longevity in multiple populations, explicitly capturing the cross-population dependence.[6] In an award-winning paper, Richman and Wüthrich also deal with multipopulation mortality modeling, extending the Lee-Carter model and using neural networks to select an optimal model structure.[7]

In the context of insurance operations, computational statistics and machine learning give us tools for modeling policyholder behavior and claims development. Hu et al use the spatial characteristics of life insurance policyholders to predict lapses by integrating census demographics with companies’ own data.[8] Kwasa and Jones apply machine learning to non-life insurance reserving, developing a support vector regression approach with a kernel function that preserves the statistical features of the loss data.[9]

Finally, this special issue introduces an expansion in the scope of the Annals to include contributions to actuarial and statistical software. As open-source software has become crucial to our research and to applications of statistical models in practice, the journal is pleased to acknowledge academics’ multifaceted contributions in that domain. Tseung et al. developed the julia package LRMoE, which enables the flexible modeling of insurance loss frequencies and severities using the Logit-weighted Reduced Mixture-of-Experts model.[10] In the same theme, Hu et al introduce the mvClaim package in R, which focuses on frameworks for multivariate insurance claim severity modeling, specifically mixtures of experts with bivariate gamma distributions and finite mixtures of copula regressions.[11] Finally, Pesenti et al. present the R package SWIM, which implements a sensitivity analysis approach that allows users to produce stressed versions of their simulation models without requiring additional model runs or full model specifications.[12]

This special issue was the result of a substantial collective effort. The editors wish to thank all the authors for their contributions and the reviewers who, with their thoughtful comments, have upheld the Annals’ rigorous standards.

A further special issue of the journal is planned. This time the theme will be Managing the Risk of Mortality Shocks, and the deadline for submissions is September 30, 2021.


Andreas Tsanakas is editor-in-chief of the Annals of Actuarial Science (AAS). He can be reached at TsanakasAAS@gmail.com.


Endnotes

[1] Antonio, Katrien, Christophe Dutang, and Andreas Tsanakas. 2020. Editorial. Annals of Actuarial Science 15, no. 2:205–206.

[2] Richman, Ronald. 2020. AI in Actuarial Science—A Review of Recent Advances—Part 1. Annals of Actuarial Science 15, no. 2:207–229; Richman, Ronald. 2020. AI in Actuarial Science—A Review of Recent Advances—Part 2. Annals of Actuarial Science 15, no. 2:230–258.

[3] Fernandez-Arjona, Lucio. 2020. A Neural Network Model for Solvency Calculations in Life Insurance. Annals of Actuarial Science 15, no. 2:259–275.

[4] Zhu, Rui, and Mario V. Wüthrich. 2020. Clustering Driving Styles via Image Processing. Annals of Actuarial Science 15, no. 2:276–290.

[5] Peters, Gareth W., Hongxuan Yan, and Jennifer Chan. Statistical Features of Persistence and Long Memory in Mortality Data. Annals of Actuarial Science 15, no. 2:291–317.

[6] Huynh, Nhan, and Mike Ludkovski. 2020. Multi-output Gaussian Processes for Multi-population Longevity Modelling. Annals of Actuarial Science 15, no. 2:318–345.

[7] Richman, Ronald, and Mario V. Wüthrich. 2020. A Neural Network Extension of the Lee-Carter Model to Multiple Populations. Annals of Actuarial Science 15, no. 2:346–366.

[8] Hu, Sen, Adrian O’Hagan, James Sweeney, and Mohammadhossein Ghahramani. 2020. A Spatial Machine Learning Model for Analysing Customers’ Lapse Behaviour in Life Insurance. Annals of Actuarial Science 15, no. 2:367–393.

[9] Kwasa, Shadrack, and Daniel Jones. 2020. A Practical Support Vector Regression Algorithm and Kernal Function for Attritional General Insurance Loss Estimation. Annals of Actuarial Science 15, no. 2:394–418.

[10] Tseung, Spark C., Andrei L. Badescu, Tsz Chai Fung, and X. Sheldon Lin. 2020. LRMoE.jl: A Software Package for Insurance Loss Modelling Using Mixture of Experts Regression Model. Annals of Actuarial Science 15, no. 2:419–440.

[11] Hu, Sen, T. Brendan Murphy, and Adrian O’Hagan. 2020. mvClaim: An R Package for Multivariate General Insurance Claims Severity Modelling. Annals of Actuarial Science 15, no. 2:441–457.

[12] Pesenti, Sylvana M., Alberto Bettini, Pietro Millossovich, and Andreas Tsanakas. 2020. Scenario Weights for Importance Measurement (SWIM)—An R Package for Sensitivity Analysis. Annals of Actuarial Science 15, no. 2:458–483.