臺灣博碩士論文加值系統

本文考慮在迴歸模型中，當依變項變數為右尾或左尾截斷缺失時的估計問題。
我們將以模擬分析比較簡易估計法、最大概似估計法與馬可夫鏈蒙地卡羅法等三
種估計方法之表現，並提出建議。除了一般的線性迴歸外，也將針對卜瓦松迴歸
模型做分析討論。文中並提供三個實例分析的結果。

We consider the estimate question of dependent variable that is rightcensored
or left-censored. And propose simulation analysis to compare
three difference estimate methods including Naive, Maximum Likelihood
Estimation (MLE), and Markov Chain Monte Carlo (MCMC). We did not only
analyze general regression model, but also discuss the Poisson regression
model. In this article, we provide the results of three empirical
analyses.

第一章 緒論................................................................................................................1
第二章 文獻探討........................................................................................................3
2.1 簡易估計法..................................................................................................4
2.2 Fair疊代法.................................................................................................4
2.3 最大概似估計法..........................................................................................6
2.4 貝氏估計法..................................................................................................7
第三章 有缺失值的卜瓦松迴歸模型......................................................................10
3.1 簡易估計法................................................................................................10
3.2 最大概似估計法........................................................................................11
3.3 貝氏估計法................................................................................................12
第四章 模擬與實例分析..........................................................................................14
4.1 實例分析....................................................................................................14
4.2 模擬方法....................................................................................................21
4.3 模擬結果分析............................................................................................22
第五章 結論..............................................................................................................24
參考文獻......................................................................................................................25
附錄 A........................................................................................................................26
附錄 B........................................................................................................................27
附錄 C........................................................................................................................43

Abbas A., Peter S. (1982) An Investigation of the Robustness of the Tobit
Estimator to Non-Normality. Econometrica, 50 50, 1055-1064.
Kai L. (1998) Bayesian Inference in a Simultaneous Equation Model with
Limited Dependent Variables. Journal of Econometrics, 85 85, 387-400.
Kishore G. (1998) Comparing Theories of Endogenous Protection: Bayesian
Comparison of Tobit Models Using Gibbs Sampling Output. The Review of
Economics and Statistics, 80 80, 128-140.
Lung F. L. (1985) The Common Structure of Tests for Selectivity Bias,
Serial Correlation, Heteroscedasticity and Non-Normality in the Tobit
Model. International Economic Review, 26 26, 1-20.
Nicky B., Mary K. C., Karen V. (1997) CODA: Convergence Diagnosis and
Output Analysis Software for Gibbs Sampling Output, Version 4.0, MRC
Biostatistics Unit, Cambridge.
Ray C. F. (1978) A Theory of Extramarital Affairs. Journal of Political
Economy, 86 86, 45-61.
Randall J. O. (1978) Note on the Uniqueness of the Maximum Likelihood
Estimator for the Tobit Model. Econometrica, 46 46, 1211-1215.
Songnian C.; Shakeeb K. (2000) Estimation of a Nonparametric Censored
Regression Model, unpublished manuscript.