Claims frequency modeling using telematics car driving data

Gao, Guangyuan; Meng, Shengwang; Wüthrich, Mario V.

doi:10.1080/03461238.2018.1523068

Cited by 59 publications

(35 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Cars could be connected to the internet of things in the near future. Then, the incumbent insurer could have access to the driver's behavior in continuous time (see Wüthrich (2017), Denuit et al (2019) and Gao et al (2019) for an analysis of this type of data). Data will then be organized as risk exposure spells (car trips).…”

Section: A3 Estimations Using the Generalized Methods Of Moments (Gmmentioning

confidence: 99%

Poisson Models With Dynamic Random Effects and Nonnegative Credibilities Per Period

Pinquet

2020

ASTIN Bull.

View full text Add to dashboard Cite

This paper provides a toolbox for the credibility analysis of frequency risks, with allowance for the seniority of claims and of risk exposure. We use Poisson models with dynamic and second-order stationary random effects that ensure nonnegative credibilities per period. We specify classes of autocovariance functions that are compatible with positive random effects and that entail nonnegative credibilities regardless of the risk exposure. Random effects with nonnegative generalized partial autocorrelations are shown to imply nonnegative credibilities. This holds for ARFIMA(0, d, 0) models. The AR(p) time series that ensure nonnegative credibilities are specified from their precision matrices. The compatibility of these semiparametric models with log-Gaussian random effects is verified. Gaussian sequences with ARFIMA(0, d, 0) specifications, which are then exponentiated entrywise, provide positive random effects that also imply nonnegative credibilities. Dynamic random effects applied to Poisson distributions are retained as products of two uncorrelated and positive components: the first is time-invariant, whereas the autocovariance function of the second vanishes at infinity and ensures nonnegative credibilities. The limit credibility is related to the three levels for the length of the memory in the random effects. The limit credibility is less than one in the short memory case, and a formula is provided.

show abstract

Section: A3 Estimations Using the Generalized Methods Of Moments (Gmmentioning

confidence: 99%

Poisson Models With Dynamic Random Effects and Nonnegative Credibilities Per Period

Pinquet

2020

ASTIN Bull.

View full text Add to dashboard Cite

show abstract

“…Predictors extrapolated from telematics data integrate traditional statistical predictors, such as age and sex of the driver or vehicle engine power, in view of the possibility of finding strong correlations between past and future (Baecke and Bocca, 2017;Guillen et al, 2019a;Wu¨thrich, 2017). Unlike traditional statistical factors, signals based on telematics data are obtained directly from the behaviour of the insured, while classic statistical data only offers proxy variables with respect to the prediction of future events (Ayuso et al, 2016;Baecke and Bocca, 2017;Denuit et al, 2019;Gao et al, 2019;Guillen et al, 2019a;Ma et al, 2018). Taking this difference into account, some research (Verbelen, 2018Wu¨thrich, 2017: 1ff) has hypothesised that telematic predictors not only work better but could even replace statistical variables in the near future offering, among other things, an effective strategy to circumvent the European legislation which prohibits the use of gender as variable in the pricing of motor insurance policies as a discriminatory practice.…”

Section: Discrimination and Fairnessmentioning

confidence: 99%

From pool to profile: Social consequences of algorithmic prediction in insurance

Cevolini

Esposito

2020

Big Data & Society

View full text Add to dashboard Cite

The use of algorithmic prediction in insurance is regarded as the beginning of a new era, because it promises to personalise insurance policies and premiums on the basis of individual behaviour and level of risk. The core idea is that the price of the policy would no longer refer to the calculated uncertainty of a pool of policyholders, with the consequence that everyone would have to pay only for her real exposure to risk. For insurance, however, uncertainty is not only a problem – shared uncertainty is a resource. The availability of individual risk information could undermine the principle of risk-pooling and risk-spreading on which insurance is based. The article examines this disruptive change first by exploring the possible consequences of the use of predictive algorithms to set insurance premiums. Will it endanger the principle of mutualisation of risks, producing new forms of discrimination and exclusion from coverage? In a second step, we analyse how the relationship between the insurer and the policyholder changes when the customer knows that the company has voluminous, and continuously updated, data about her real behaviour.

show abstract

“…The prominent actuarial journals listed in Appendix A were searched for the key phrase “neural networks,” and the search was augmented by similar searches on the SSRN and arxiv repositories. Articles making only a passing reference to neural networks were dropped from the list, and the resulting list of topics and papers reviewed in this section is as follows: pricing of non-life insurance (Noll et al , 2018; Wüthrich & Buser, 2018); incurred but not reported (IBNR) reserving (Kuo, 2018b; Wüthrich, 2018a; Zarkadoulas, 2017) and the individual claims simulation machine of Gabrielli and Wüthrich (2018); analysis of telematics data (Gao et al , 2018; Gao & Wüthrich, 2017; Wüthrich & Buser, 2018; Wüthrich, 2017); mortality forecasting (Hainaut, 2018a); approximating nested stochastic simulations (Hejazi & Jackson, 2016, 2017); forecasting financial markets (Smith et al , 2016). …”

Section: Survey Of Deep Learning In Actuarial Sciencementioning

confidence: 99%

AI in actuarial science – a review of recent advances – part 2

Richman

2020

Ann. actuar. sci.

View full text Add to dashboard Cite

Rapid advances in artificial intelligence (AI) and machine learning are creating products and services with the potential not only to change the environment in which actuaries operate, but also to provide new opportunities within actuarial science. These advances are based on a modern approach to designing, fitting and applying neural networks, generally referred to as “Deep Learning”. This paper investigates how actuarial science may adapt and evolve in the coming years to incorporate these new techniques and methodologies. Part 1 of this paper provides background on machine learning and deep learning, as well as an heuristic for where actuaries might benefit from applying these techniques. Part 2 of the paper then surveys emerging applications of AI in actuarial science, with examples from mortality modelling, claims reserving, non-life pricing and telematics. For some of the examples, code has been provided on GitHub so that the interested reader can experiment with these techniques for themselves. Part 2 concludes with an outlook on the potential for actuaries to integrate deep learning into their activities. Finally, a supplementary appendix discusses further resources providing more in-depth background on machine learning and deep learning.

show abstract

Claims frequency modeling using telematics car driving data

Cited by 59 publications

References 8 publications

Poisson Models With Dynamic Random Effects and Nonnegative Credibilities Per Period

Poisson Models With Dynamic Random Effects and Nonnegative Credibilities Per Period

From pool to profile: Social consequences of algorithmic prediction in insurance

AI in actuarial science – a review of recent advances – part 2

Contact Info

Product

Resources

About