臺灣博碩士論文加值系統

In this article, robust regression parameter inference in the setting of generalized linear models (GLMs) will be proposed. A parametric robust regression methodology that is robust to violations of the distributional assumptions is able to test hypotheses on the regression coefficients in the misspecified GLM setting. More specifically, it will be demonstrated that with large samples the ordinary normal, gamma and inverse Gaussian regression models can be made robust and provide consistent regression parameter estimates in the misspecified GLM setting. These adjusted regression models furnish the correct type I, II error probabilities, and also the correct coverage probability, for continuous data, as long as the true but unknown underlying distributions have finite second moments.

The parametric robust regression techniques are also applied to the analysis of variance (ANOVA) problems including the one-way, two-way ANOVA structures and the one-way analysis of covariance (ANCOVA) setup. In the ANOVA situations, these adjusted regression models continue to remain asymptotically valid representations of the particular parameters of interest, whatever distributions generate the data.

Abstract I
Acknowledgments II
Contents III
1 Introduction 1
2 Robust Tests Based on the Adjusted Likelihood Function 4
2.1 Model Msspecification …………………………………………………... 4
2.2 Robust Likelihood ……………………………………………………….. 11
2.2.1 Models with a One-Dimensional Parameter …………………….. 11
2.2.2 Models with Nuisance Parameters ………………………………. 14
3 Robust Likelihood Inference for the Regression Parameters in the
Setting of Generalized Linear Models 20
3.1 The Large-Sample Behavior of the Maximum Likelihood (ML)
Estimator in Misspecified Generalized Linear Models …………………. 20
3.1.1 Misspecified Generalized Linear Models ……………………….. 21
3.1.2 Asymptotic Properties of the ML Estimate ……………………... 26
3.2 Asymptotic Behavior of the ML Estimate Based on the Superior Working
Models ………………………………………………………………….. 29
3.3 Making Regression Models Robust ……………………………………... 41
4 Robust Regression Models 47
4.1 Robust Normal Regression ………………………………………………. 47
4.2 Robust Gamma Regression ……………………………………………… 50
4.3 Robust Inverse Gaussian Regression …………………………………….. 52
5 Analysis of Variance Models 55
5.1 One-Way ANOVA ………………………………………………………. 55
5.1.1 Normal Working Models ………………………………………... 56
5.1.2 Gamma Working Models ……………………………………….. 61
5.1.3 Inverse Gaussian Working Models ……………………………… 64
5.2 Two-Way ANOVA ………………………………………………………. 67
5.2.1 Normal Working Models ………………………………………... 68
5.2.2 Gamma Working Models ……………………………………….. 72
5.2.3 Inverse Gaussian Working Models ……………………………… 74
5.3 One-Way ANCOVA ……………………………………………………... 76
5.3.1 Normal Working Models ………………………………………... 77
5.3.2 Gamma Working Models ……………………………………….. 78
5.3.3 Inverse Gaussian Working Models ……………………………… 79
6 Simulation Studies 81
6.1 Regression Models ………………………………………………………. 81
6.2 Analysis of Variance Models ……………………………………………. 89
6.2.1 One-Way ANOVA ……………………………………………… 89
6.2.2 Two-Way ANOVA ……………………………………………… 103
6.2.3 One-Way ANCOVA …………………………………………….. 110
7 A Real Example 117
8 Concluding Remarks 119
References 121
Appendix A Adjusting Matrices 124
Appendix B Adjusted Covariance Matrices 132

Birnbaum, A. (1962). On the Foundations of Statistical Inference (with Discussion). J. Amer. Statist. Assoc., 53, 259-326.
Cox, D. R. and Hinkley, D. V. (1974). Theoretical Statistics. London: Chapman and Hall.
Drygas, H. (1976). Weak and Strong Consistency of the Least Squares Estimators in Regression Models. Z. Wahrsch. Verw. Gebiete, 34, 119-127.
Fahrmeir, L. and Kaufmann, H. (1985). Consistency and Asymptotic Normality of the Maximum Likelihood Estimator in Generalized Linear Models. The Annals of Statistics, 13, 342-368.
Fahrmeir, L. and Kaufmann, H. (1986). Asymptotic Inference in Discrete Response Models. Statistical Papers, 27, 179-205.
Fahrmeir, L. (1987). Asymptotic Testing Theory for Generalized Linear Models. Statistics, 18, 65-76.
Fahrmeir, L. (1990). Maximum Likelihood Estimation in Misspecified Generalized Linear Models. Statistics, 21, 487-502.
Fahrmeir, L. and Tutz, G. (2001). Multivariate Statistical Modelling Based on Generalized Linear Models. 2nd ed. New York: Springer-Verlag.
Gourieroux, C., Monfort, A. and Trognon, A. (1984). Pseudo Maximum Likelihood Methods: Theory. Econometrica, 52, 681-700.
Hall, A. R. and Inoue, A. (2003). The Large Sample Behavior of the Generalized Method of Moments Estimator in Misspecified Models. Journal of Econometrics, 114, 361-394.
Heagerty, P. J. and Kurland, B. F. (2001). Misspecified Maximum Likelihood Estimates and Generalised Linear Mixed Models. Biometrika, 88, 973-985.
Helsel, D. R. (1983). Mine Drainage and Rock Type Influences on Eastern Ohio Street Water Quality. Water Resources Bulletin, 19, 881-887.
Heyde, C. C. (1997). Quasi-Likelihood and Its Application. New York: Springer-Verlag.
Huber, P. J. (1967). The Behavior of Maximum Likelihood Estimates under Nonstandard Conditions. Proc. Fifth Berkeley Symp. Math. Statist. Probab., 1, 221-233. Univ. California Press, Berkeley.
Huber, P. J. (1981). Robust Statistics. New York: John Wiley.
Kaufmann, H. (1988a). On Existence and Uniqueness of a Vector Minimizing a Convex Function. ZOR, 32, 357-373.
Kaufmann, H. (1988b). On Existence and Uniqueness of Maximum Likelihood Estimates in Quantal and Ordinal Response Models. Metrika, 35, 291-313.
Kent, T. J. (1982). Robust Properties of Likelihood Ratio Tests. Biometrika, 69, 19-27.
Lai, T. L., Robbins, H. and Wei, C. Z. (1979). Strong Consistency of Least Squares Estimates in Multiple Regression II. J. Multivariate Anal., 9, 343-361.
Lehmann, E. L. (1983). Theory of Point Estimation. New York: Springer-Verlag.
Lehmann, E. L. (1999). Elements of Large-Sample Theory. New York: Springer-Verlag.
Mardia, K. V., Kent, J. T. and Bibby, J. M. (1979). Multivariate Analysis. London, New York: Academic Press.
McCullagh, P. (1983). Quasi-Likelihood Functions. Annals of Statistics, 11, 59-67.
McCullagh, P. and Nelder, J. A. (1989). Generalized Linear Models. 2nd ed. New York: Chapman and Hall.
Nelder, J. A. and Wedderburn, R. W. M. (1972). Generalized Linear Models. JRSS-A, 135, 370-384.
Rao, C. R. and Toutenburg, H. (1999). Linear Models – Least Squares and Alternatives. 2nd ed. New York: Springer-Verlag.
Rawlings, J. O., Pantula, S. G. and Dickey, D. A. (1998). Applied Regression Analysis – A Research Tool. 2nd ed. Now York: Springer-Verlag.
Rohatgi, V. K. (1976). An Introduction to Probability Theory and Mathematical Statistics. New York: John Wiley.
Royall, R. M. (1997). Statistical Evidence – A Likelihood Paradigm. London: Chapman and Hall.
Royall, R. M. (2000). On the Probability of Observing Misleading Statistical Evidence (with Discussion). J. Amer. Statist. Assoc., 95, 760-780.
Royall, R. M. and Tsou, T.-S. (2003). Interpreting Statistical Evidence Using Imperfect Models: Robust Adjusted Likelihood Functions. JRSS-B, 65, 391-404.
Stafford, J. E. (1996). A Robust Adjustment of the Profile Likelihood. The Annals of Statistics, 24, 336-352.
Tsou, T.-S. (2003). Comparing Two Population Means and Variances – A Parametric Robust Way. Comm. Stat. – Theor. Meth., 32, 2013-2019.
Tsou, T.-S. and Cheng, K.-F. (2004). Parametric Robust Regression Analysis of Contaminated Data. Comm. Stat. – Theor. Meth., 33, 1887-1898.
Tsou, T.-S. (2004). Parametric Robust Inferences for Regression Parameters under Generalized Linear Models. (submitted)
Tsou, T.-S. (2005). Robust Poisson Regression. J. Statist. Plan. Infer. (to appear)
Wedderburn, R. W. M. (1974). Quasi-Likelihood Functions, Generalized Linear Models, and the Gauss-Newton Method. Biometrika, 61, 439-447.
Wedderburn, R. W. M. (1976). On the Existence and Uniqueness of the Maximum Likelihood Estimates for Certain Generalized Linear Models. Biometrika, 63, 27-32.
White, H. (1982). Maximum Likelihood Estimation of Misspecified Models. Econometrica, 50, 1-25.
White, H. and Domowitz, I. (1984). Nonlinear Regression with Dependent Observations. Econometrica, 52, 143-161.
White, H. (1984). Maximum Likelihood Estimation of Misspecified Dynamic Models. In: Dijlestra, T. (Ed.), Misspecification Analysis. Berlin: Springer-Verlag.