透過您的圖書館登入
IP:18.191.180.252
  • 學位論文

利用中位數插補法分析設限資料之線性迴歸模型

The Median Imputation Approach for Linear Regression Model under Censored Data

指導教授 : 謝進見
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


在本篇論文中,我們探討設限資料的線性迴歸模型。 我們想要比較Buckley-James (1979)的平均數差補法和中位數差補法之間的差異性。 以及藉由這兩種方法去差補設限資料和利用最小平方估計法的疊代過程去估計出迴歸係數。 然後使用bootstrap method去估計變異數。 針對這兩種方法進行大量的模擬研究,比較其在有限樣本之下的特性。 最後,我們將利用這兩種方法分析肺癌資料的例子(Loprinzi et al., 1994)。

並列摘要


In this thesis, we consider the linear regression model for censored data. We would like to compare the difference between the mean imputation approach by Buckley-James (1979) and the median imputation approach. We would consider the two approaches to impute the censored data and estimate the regression parameters with LSE iterative procedure. Then, we estimate the variance by bootstrap method. We compare the finite-sample performance of the two approaches via simulation studies. Finally, we consider the Lung Cancer data by Loprinzi et al. (1994) for illustrations.

參考文獻


naires., North Central Cancer Treatment Group. Journal of Clinical
New York: Chapman and Hall.
[3] Buckley, J. and James I. (1979) Linear Regression with Censored Data.,
[4] Miller, P. (1976) Least Squares Regression with Censored Data.,
Biometrika 63, 449-64.

延伸閱讀