透過您的圖書館登入
IP:18.118.120.204
  • 學位論文

Cox比例風險模型之參數估計與比例風險檢定-比較部分概似法、二階段方法以及聯合模型法

Proportional Hazards Test and Estimation for Cox Proportional Hazards model ---- Among Partial Likelihood, Two-stage and Joint modeling approach

指導教授 : 曾議寬
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


在生物醫學的研究的過程中,有興趣的是時間相依共變量與存活時間的關聯性,而推估關聯性最常使用的是Cox比例風險迴歸模型。傳統上使用Cox(1972)的部分概似法估計參數,但前提是必須有所有研究對象的完整共變量資訊且不允許誤差。為了減少部分概似法對於有遺失值時其參數估計上的偏差,本研究分別採用二階段方法以及聯合模型(Wulfsohn,1997)來估計參數,目的是比較此三種方法之下,對於不同遺失比例其參數估計值變化情形,以及通過比例風險假設之比例,以利於不同條件之下選擇最有效之方法,模擬結果在不同的共變量軌跡之下,若無遺失比例之發生則可選擇程式效率較高之部分概似法,在有遺失比例發生但測量誤差不大時可選擇二階段方法,若有遺失比例之發生且測量誤差較高時,則需使用聯合模型來估計參數,其結果較二階段模型快速且準確。

並列摘要


The relationship between longitudinal covariates and a failure time process can be assessed using the Cox proportional hazards model. The purpose of the study is to evaluate the performance of three approaches, the partial likelihood, two-stage partial likelihood , and joint model approaches for the Cox model when the covariate is measured in irregular times with measurements error. The results show that partial likelihood is an efficient choice when data has no missing values; two-stage model can be selected for data with small measure error. As a conclusion, joint modeling approach is the best choice in all situations.

參考文獻


[1] Anderson, P. K. and Gill, R. D. (1982). Cox’s regression model forcounting processes, a large sample study. Annals of Statistics 10.1100-1120.
[2] Burden, Richard L. and Faires, J. Douglas (2000). Numerical Analysis (7th ed.). Brooks/Cole
[4] Cox, D. R. (1975). Partial likelihood. Biometrika 62. 269-276.
[5] Efron, B. (1994). Missing data, imputation and bootstrap (with Discussion). J. Am. Statist. Assoc. 89. 463-479.
[6] Hsieh, F., Tseng, Y. K., and Wang, J. L. (2006). Joint modelingof survival time and longitudinal data: likelihood approach revisit.Biometrics, 62.1037-1043.

延伸閱讀