加速壽命模型中自變數有重複測量時之估計方法

The estimation of accelerated failure time model when there are replicate measurements of covariates

陳立誼(Li-Yi Chen)

指導教授 ： 黃逸輝

淡江大學/理學院/數學學系碩士班/碩士(2012年)

https://doi.org/10.6846/TKU.2012.01330

摘要

加速壽命模型(AFT model , accelerated failure time model)具有相當直覺性的解釋，因此它是除了 Cox 比例風險模型外另一個常用的統計模型。但當自變數受測量誤差影響時，目前還未能提出具一致性的估計方法。本文考慮在自變數有重複測量的情形下，修正Tsiatis(1990)提出的估計方法以期能找出在自變數受測量誤差影響能具一致性的估計方法。而我們提出的估計方法具一致性的推測不僅可從理論來證明也可從電腦模擬結果中得到驗證。

關鍵字

加速壽命模型 ； 測量誤差 ； 重複測量

並列摘要

Accelerated failure time model is an attractive alternative to the Cox proportional hazard model since it is more intuitive in interpretation. When the covariates are measured with errors, there is no consistent estimation method so far. When there are replicate measurements, we modified the estimation in Tsiatis (1990) to account for measurement errors. Our method is expected to be consistent and this conjecture is supported by our simulation results.

並列關鍵字

measurement error ； accelerated failure time model ； replicate measurements

參考文獻

failure time models with covariates subject to measurement

Huang, Y. (2002). Calibration Regression of Censored

Rank-based inference for the accelerated failure time

model. Biometrika 90, 341-353.

failure time models for counting processes. Biometrika