Median bilinear models in presence of extreme values. (English) Zbl 1482.62061

Santolino, Miguel

SORT 45, No. 2, 163-180 (2021).

Summary: Bilinear regression models involving a nonlinear interaction term are applied in many fields (e.g., Goodman’s RC model, Lee-Carter mortality model or CAPM financial model). In many of these contexts data often exhibit extreme values. We propose the use of bilinear models to estimate the median of the conditional distribution in the presence of extreme values. The aim of this paper is to provide alternative methods to estimate median bilinear models. A calibration strategy based on an iterative estimation process of a sequence of median linear regression is developed. Mean and median bilinear models are compared in two applications with extreme observations. The first application deals with simulated data. The second application refers to Spanish mortality data involving years with atypical high mortality (Spanish flu, civil war and HIV/AIDS). The performance of the median bilinear model was superior to that of the mean bilinear model. Median bilinear models may be a good alternative to mean bilinear models in the presence of extreme values when the centre of the conditional distribution is of interest.

MSC:

62H12	Estimation in multivariate analysis
62H17	Contingency tables
62L12	Sequential estimation

Keywords:

outliers; quantile regression; single factor models; nonlinear; multiplicative

Software:

R; gnm; quantreg; Human Mortality; GitHub; Median-Bilinear-Models; mss

References:

[2]	Anderson, J.A. (1984). Regression and ordered categorical variables (with Discussion). Journal of the Royal Statistical Society. Series B (Methodological), 47, 203-210.
[3]	Atance, D., Debón, A., and Navarro, E. (2020) A comparison of forecasting mortality models using resampling methods. Mathematics, 8, 1550.
[4]	Bassett, G. and Koenker, R. (1978). Theory of least absolute error regression. Journal of the American Statistical Association, 73, 618-622. · Zbl 0391.62046
[5]	Black, F., Jensen, M.C. and Scholes, M. (1972). The capital asset pricing model: some empirical tests. Studies in the theory of capital markets, New York, Praeger: 79-121.
[6]	Brouhns, N., Denuit, M., and Vermunt, J. (2002). A Poisson log-bilinear regression approach to the construction of projected life table. Insurance: Mathematics and Economics, 31, 373-393. · Zbl 1074.62524
[7]	Buchinsky, M. (1995). Estimating the asymptotic covariance matrix for quantile regres-sion models a Monte Carlo study. Journal of Econometrics, 68, 303-338. · Zbl 0825.62437
[8]	Carreras, A. and Tafunell, X.(2005). Estadísticas Hist´ na: siglos XIX-XX. oricas de Espa˜ Fundación BBVA, 2a Edic., Bilbao. ´
[9]	CNE (2011). Area de vigilancia de VIH y conductas de riesgo. Mortalidad por VIH/Sida en Espa˜ no 2009. Evoluci´ Centro Nacional de Epimediología, na, a˜ on 1981-2009.
[10]	Secretaría del Plan Nacional Sobre el Sida, Gobierno de Espaã.
[11]	Croux, C., Filzmoser, P., Pison, G. and Rousseeuw, P.J. (2003). Fitting multiplicative models by robust alternating regressions. Statistics and Computing , 13, 23-36.
[12]	Denis, J.B. and Pázman, A.(1999). Bias of LS estimators in nonlinear regression models with constraints. Part II: Biadditive models. Applications of Mathematics, 44, 375-403. · Zbl 1060.62527
[13]	Dutang, C. (2017). Some explanations about the IWLS algorithm to ft generalized linear models. Technical report, hal-01577698.
[14]	El-Attar, R. A., Vidyasagar, M., and Dutta, S. R. K. (1979). An Algorithm for l 1 -norm minimization with application to nonlinear l 1 -approximation. SIAM Journal on Nu-merical Analysis, 16, 70-86. · Zbl 0401.90089
[15]	Emerson, J.D. and Hoaglin. D.C. (1983). Analysis of two-way tables by medians. In D. C. Hoaglin, F. Mosteller and J. W. Tukey (Eds.), Understanding Robust and Ex-ploratory Data Analysis, 165-210, New York City: John Wiley and Sons. · Zbl 0599.62007
[16]	Erikson, R. and Goldthorpe, J.H. (1992). The Constant Flux: A Study of Class Mobility in Industrial Societies. Oxford: Clarendon Press. .Ch. 3.
[17]	FMBM (2021). Fitting Median Bilinear Model. Available at https://github.com/ msantolino/Median-Bilinear-Models. Accessed 8 November 2021.
[18]	Gabriel, K.R. (1978). Least squares approximation of matrices by additive and multi-plicative models. Journal of the Royal Statistical Society. Series B (Methodological), 40, 186-196. · Zbl 0393.62019
[19]	Gabriel, K. R. and Odoroff, L. (1984). Resistant lower rank approximation of matrices. In E. Diday, M. Jambu, L. Lebart, J. Pages and R. Tomassone (Eds.), Data analysis and informatics III, 23-40, Amsterdam: North-Holland.
[20]	Goodman, L.A. (1979). Simple models for the analysis of association in cross-classifcations having ordered categories. Journal of the American Statistical Asso-ciation, 74, 537-552.
[21]	Goodman, L.A. (1981). Association models and canonical correlation in the analysis of cross-classifcations having ordered categories. Journal of the American Statistical Association, 76, 320-334.
[22]	Hawkins, D.M. (1980). Identifcation of outliers. Monographs on applied probability and statistics, Chapman & Hall.
[23]	HMD (2020). Human Mortality Database. University of California, Berkeley (USA), and Max Planck Institute for Demographic Research (Germany). Available at www. mortality.org. University of California, Accessed 15 January 2020.
[24]	Jiménez Lucena, I. (1994). El tifus exantemático de la posguerra española (1939-1943): el uso de una enfermedad colectiva en la legitimación del ’Nuevo estado’. Dynamis : Acta Hispanica ad Medicinae Scientiarumque. Historiam Illustrandam, 14, 185-198.
[25]	Justel, A., Peña, D. and Tay, R.S (2001). Detection of outlier patches in autoregressive time series. Statistica Sinica, 11, 651-673. · Zbl 0978.62081
[26]	Koenker, R. (2019). quantreg: Quantile regression. R package version 5.42.
[27]	Koenker, R. (2020). Non linear quantile regression. http://www.econ.uiuc.edu/ ∼ roger/ research/nlrq/nlrq.html, Accessed 16 February 2021.
[28]	Koenker, R. and Park, B. J. (1996). An interior point algorithm for nonlinear quantile regression. Journal of Econometrics, 71, 265-283. · Zbl 0855.62030
[29]	Lee, R. D. and Carter, L. R. (1992). Modeling and forecasting U. S. mortality. Journal of the American Statistical Association, 87, 659-671.
[30]	Machado, J. and Silva, J. S. (2011). MSS: Stata module to perform heteroskedastic-ity test for quantile and OLS regressions. Statistical Software Components, Boston College Department of Economics.
[31]	Macias, Y. and Santolino, M. (2018). Application of Lee-Carter and Renshaw-Haberman models in life insurance products. Anales del Instituto de Actuarios Españoles, 24, 53-78.
[32]	Mehrotra, S. (1992). On the implementation of a primal−dual interior point method. SIAM Journal on Optimization, 2, 575-601. · Zbl 0773.90047
[33]	Moyano-Silva, P.A., Perez-Maŕ ín, A.M. and Santolino, M. (2020). Estimation of stochastic mortality models for Chile. Anales del Instituto de Actuarios Españoles, 4, 225-256.
[34]	Portnoy, S. and Koenker, R. (1997). The Gaussian hare and the Laplacian tortoise: computability of squared-error versus absolute-error estimators. Statistical Science, 12, 279-300. · Zbl 0955.62608
[35]	R Core Team, (2020) R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna.
[36]	Renshaw, A. and Haberman, S. (2006). A cohort-based extension to the Lee-Carter model for mortality Reduction Factors. Insurance: Mathematics and Economics, 38:556-570. · Zbl 1168.91418
[37]	Robine, J.M, Crimmins, E.M, Horiuchi S. and Zeng Yi, Z.(2007). Human Longevity, Individual Life Duration, and the Growth of the Oldest-Old Population. International Studies in Population, Springer.
[38]	Salahi, M., Peng, J. and Terlaky, T. (2008) On Mehrotra-type predictor-corrector algo-rithms. SIAM Journal on Optimization, 18, 1377-1397. · Zbl 1165.90569
[39]	Sánchez, B., Labros, H. and Labra, V. (2013). Likelihood based inference for quantile regression using the asymmetric Laplace distribution. Journal of Statistical Compu-tation and Simulation, 81, 1565-1578
[40]	Santolino, M. (2020). The Lee-Carter quantile mortality model. Scandinavian Actuarial Journal, 7, 614-633 · Zbl 1448.91265
[41]	Turner, H. and Firth, D. (2018). Generalized nonlinear models in R: an overview of the gnm package. R package version 1.1-0.
[42]	Turner, H., Firth, D. and Kosmidis, I. (2013). Generalized nonlinear models in R. 6th International Conference of the ERCIM WG on Computational and Methodological Statistics, ERCIM, London.
[43]	Van Eeuwijk, Fred A. (1992). Multiplicative models for genotype-environment interac-tion in plant breeding. Statistica Applicata, 4, 393-406.
[44]	Van Eeuwijk, Fred A. (1995). Multiplicative interaction in generalized linear models. Biometrics , 51, 1017-032. · Zbl 0875.62330
[45]	Wilmoth, J. (1993). Computational Methods for Fitting and Extrapolating the Lee-Carter Model of Mortality Change. Technical Report. Department of Demography. University of California.
[46]	Xie, Y. (1992). The log-multiplicative layer effect model for comparing mobility tables. American Sociological Review, 57, 380-395.
[47]	Yates, F. and Cochran, W.G. (1938). The analysis of groups of experiments. The Journal of Agricultural Science, 28, 556-580