Minimum Density Power Divergence Estimator for Negative Binomial Integer-Valued GARCH Models

Abstract

In this paper, we study a robust estimation method for the observation-driven integer-valued time-series models in which the conditional probability mass of current observations is assumed to follow a negative binomial distribution. Maximum likelihood estimator is highly affected by the outliers. We resort to the minimum density power divergence estimator as a robust estimator and show that it is strongly consistent and asymptotically normal under some regularity conditions. Simulation results are provided to illustrate the performance of the estimator. An application is performed on data for campylobacteriosis infections.

Keywords

Integer-valued GARCH model / Minimum density power divergence estimator / Negative binomial distribution / Robust estimation

References

[1] Aeberhard WH, Cantoni E, Heritier S. Robust inference in the negative binomial regression model with an application to falls data. Biometrics. 2014, 70 920-931

[2] Basu A, Harris IR, Hjort NL, Jones MC. Robust and efficient estimation by minimising a density power divergence. Biometrika. 1998, 85 549-559

[3] Billingsley P. The Lindeberg–Lévy theorem for martingales. Proc. Am. Math. Soc.. 1961, 12 788-792

[4] Cantoni E, Ronchetti EM. Robust inference for generalized linear models. J. Am. Stat. Assoc.. 2001, 96 1022-1030

[5] Chen CWS, Lee S. Generalized Poisson autoregressive models for time series of counts. Comput. Stat. Data Anal.. 2016, 99 51-67

[6] Chen H, Li Q, Zhu F. Binomial AR(1) processes with innovational outliers. Commun. Stat. Theory Methods. 2020

[7] Christou V, Fokianos K. Quasi-likelihood inference for negative binomial time series models. J. Time Ser. Anal.. 2014, 35 55-78

[8] Christou V, Fokianos K. On count time series prediction. J. Stat. Comput. Simul.. 2015, 85 357-373

[9] Cui Y, Zheng Q. Conditional maximum likelihood estimation for a class of observation-driven time series models for count data. Stat. Probab. Lett.. 2017, 123 193-201

[10] Czado C, Gneiting T, Held L. Predictive model assessment for count data. Biometrics. 2009, 65 1254-1261

[11] Elsaied H, Fried R. Robust fitting of INARCH models. J. Time Ser. Anal.. 2014, 35 517-535

[12] Ferland R, Latour A, Qraichi D. Integer-valued GARCH process. J. Time Ser. Anal.. 2006, 27 923-942

[13] Fokianos K, Rahbek A, Tjøstheim D. Poisson autoregression. J. Am. Stat. Assoc.. 2009, 104 1430-1439

[14] Francq C, Zakoïan JM. Maximum likelihood estimation of pure GARCH and ARMA–GARCH processes. Bernoulli. 2004, 10 605-637

[15] Fried R, Agueusop I, Bornkamp B, Fokianos K, Fruth J, Ickstadt K. Retrospective Bayesian outlier detection in INGARCH series. Stat. Comput.. 2015, 25 365-374

[16] Kang J, Lee S. Minimum density power divergence estimator for Poisson autoregressive models. Comput. Stat. Data Anal.. 2014, 80 44-56

[17] Kim B, Lee S. Robust estimation for zero-inflated poisson autoregressive models based on density power divergence. J. Stat. Comput. Simul.. 2017, 87 2981-2996

[18] Kim B, Lee S. Robust estimation for general integer-valued time series models. Ann. Inst. Stat. Math.. 2020

[19] Kitromilidou K, Fokianos K. Mallows’ quasi-likelihood estimation for log-linear Poisson autoregressions. Stat. Inference Stoch. Process.. 2016, 19 337-361

[20] Li Q, Lian H, Zhu F. Robust closed-form estimators for the integer-valued GARCH(1,1) model. Comput. Stat. Data Anal.. 2016, 101 209-225

[21] Liu M, Li Q, Zhu F. Threshold negative binomial autoregressive model. Statistics. 2019, 53 1-25

[22] Liu Z, Li Q, Zhu F. Random environment binomial thinning integer-valued autoregressive process with Poisson or geometric marginal. Braz. J. Probab. Stat.. 2020

[23] Maji A, Ghosh A, Basu A, Pardo L. Robust statistical inference based on the $C$-divergence family. Ann. Inst. Stat. Math.. 2019, 71 1289-1322

[24] Marazzi A, Valdora M, Yohai V, Amiguet M. A robust conditional maximum likelihood estimator for generalized linear models with a dispersion parameter. TEST. 2019, 28 223-241

[25] Qi X, Li Q, Zhu F. Modeling time series of count with excess zeros and ones based on INAR(1) model with zero-and-one inflated Poisson innovations. J. Comput. Appl. Math.. 2019, 346 572-590

[26] Scotto MG, Weiß CH, Gouveia S. Thinning-based models in the analysis of integer-valued time series: a review. Stat. Model.. 2015, 15 590-618

[27] Straumann D, Mikosch T. Quasi-maximum-likelihood estimation in conditionally heteroscedastic time series: a stochastic recurrence equations approach. Ann. Stat.. 2006, 34 2449-2495

[28] Toma A, Broniatowski M. Dual divergence estimators and tests: robustness results. J. Multivar. Anal.. 2011, 102 20-36

[29] Valdora M, Yohai VJ. Robust estimators for generalized linear models. J. Stat. Plan. Inference. 2014, 146 31-48

[30] Xiong L, Zhu F. Robust quasi-likelihood estimation for the negative binomial integer-valued GARCH(1,1) model with an application to transaction counts. J. Stat. Plan. Inference. 2019, 203 178-198

[31] Yu M, Wang D, Yang K, Liu Y. Bivariate first-order random coefficient integer-valued autoregressive processes. J. Stat. Plan. Inference. 2020, 204 153-176

[32] Zhang C, Guo X, Cheng C, Zhang Z. Robust-BD estimation and inference for varying dimensional general linear models. Stat. Sin.. 2014, 24 653-673

[33] Zhu F. A negative binomial integer-valued GARCH model. J. Time Ser. Anal.. 2011, 32 54-67

[34] Zhu F. Modeling overdispersed or underdispersed count data with generalized Poisson integervalued GARCH models. J. Stat. Plan. Inference. 2012, 389 58-71

[35] Zhu F. Zero-inflated Poisson and negative binomial integer-valued GARCH models. J. Math. Anal. Appl.. 2012, 142 826-839

[36] Zhu F, Liu S, Shi L. Local influence analysis for Poisson autoregression with an application to stock transaction data. Stat. Neerl.. 2016, 70 4-25

Funding

National Natural Science Foundation of China(11871027)