×

Modeling hidden exposures in claim severity via the EM algorithm. (English) Zbl 1085.62515

Summary: We consider the issue of modeling the latent or hidden exposure occurring through either incomplete data or an unobserved underlying risk factor. We use the celebrated expectation maximization (EM) algorithm as a convenient tool in detecting latent (unobserved) risks in finite mixture models of claim severity and in problems where data imputation is needed. We provide examples of applicability of the methodology based on real-life auto injury claim data and compare, when possible, the accuracy of our methods with that of standard techniques. Sample data and an EM algorithm program are included to allow readers to experiment with the EM methodology themselves.

MSC:

62P05 Applications of statistics to actuarial sciences and financial mathematics
91B30 Risk theory, insurance (MSC2010)
PDF BibTeX XML Cite
Full Text: DOI

References:

[1] Abrahamse Alan F., Automobile Insurance: Road Safety, New Drivers, Risks, Insurance Fraud and Regulation pp 131– (1999)
[2] Bailey Robert A., ASTIN Bulletin 1 (4) pp 192– (1960)
[3] Bilmes Jeff A., Gentle Tutorial of the EM Algorithm and Its Application to Parameter Estimation for Gaussian Mixture and Hidden Markov Models (1998)
[4] BolancĂ© Catalina, Insurance: Mathematics and Economics 32 (1) pp 19– (2003) · Zbl 1024.62041
[5] Brockett Patrick L., Journal of Risk and Insurance 69 (3) pp 341– (2002)
[6] Brockett Patrick L., Journal of Risk and Insurance 65 (2) pp 245– (1998)
[7] Craiu Radu V., Biometrika (2004)
[8] Crocker Keith J., Automobile Insurance: Road Safety, New Drivers, Risks, Insurance Fraud and Regulation pp 119– (1999)
[9] Cummins J. David, Journal of Risk and Insurance 50 pp 377– (1983)
[10] Cummins J. David, ASTIN Bulletin 14 (1) pp 45– (1985)
[11] Cummins J. David, Journal of Economic Perspectives 6 (2) pp 95– (1996)
[12] D’arcy Stephen P., Contingencies pp 40– (1997)
[13] Dempster Allan P., Journal of Royal Statistical Society Series B 39 (1) pp 1– (1977)
[14] Derrig Richard A., Journal of Risk and Insurance 69 (3) pp 271– (2002)
[15] Derrig Richard A., Proceedings of the Casualty Actuarial Society 87 pp 322– (2001)
[16] Derrig Richard A., Journal of Risk and Insurance 61 (2) pp 245– (1994)
[17] Dugas Charles D., Intelligent Techniques for the Insurance Industry: Theory and Applications pp 97– (2003)
[18] Embrechts Paul, Extremes and Integrated Risk Management pp 71– (2000)
[19] Hagenaars Jacques, Applied Latent Class Analysis (2001)
[20] Hastie Trevor, The Elements of Statistical Learning (2001) · Zbl 0973.62007
[21] Nicholas J. Horton, American Statistician 55 (3) pp 244– (2001) · Zbl 05680456
[22] Kalb Guyonne R. J., Insurance: Mathematics & Economics 18 pp 89– (1996) · Zbl 0864.62071
[23] Little Roderick J. A., Statistical Analysis with Missing Data (1987) · Zbl 0665.62004
[24] Major John A., International Journal of Intelligent Systems 7 pp 687– (1992)
[25] Mccutcheon Allan C., Latent Class Analysis (1987)
[26] Mclachlan Geoffrey, The EM Algorithm and Extensions (1997) · Zbl 0882.62012
[27] Mclachlan Geoffrey, Finite Mixture Models (2000)
[28] Owen Art B., Empirical Likelihood (2001)
[29] Rempala Grzegorz A., Scandinavian Actuarial Journal pp 53– (2004)
[30] Rubin Donald B., Multiple Imputation for Nonresponse in Surveys (1987) · Zbl 1070.62007
[31] Rubin Donald B., Journal of the American Statistical Association 91 pp 473– (1996)
[32] Schafer Joe L., Analysis of Incomplete Multivariate Data (1997) · Zbl 0997.62510
[33] Wang Shaun, Proceedings of the Casualty Actuarial Society 83 pp 848– (1998)
[34] Wu Jeff A., Annals of Statistics 11 (1) pp 95– (1983) · Zbl 0517.62035
[35] Xu Li, Neural Computation 8 pp 129– (1996) · Zbl 05476078
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. It attempts to reflect the references listed in the original paper as accurately as possible without claiming the completeness or perfect precision of the matching.