右設限存活資料分析中的參數化加速失敗時間模型，經常假設其
存活分布為韋伯分布、對數常態分布或對數邏輯司分布，當共變數存
在時，又多展示為對數線性模式。本文考慮的參數化加速失敗時間模
式之存活分布為廣義伽瑪分布，因為此一分布包含韋伯分布、對數常
態分布及伽瑪分布等，所能涵蓋的右偏分布較為廣泛。此外，本文利
用區域概似估計(local likelihood estimation)研究對數線性模式
中共變數的可能轉換。之後本文用模擬研究本文所提比較廣義的模型
在各種不同加速失敗時間模型及存活分布下，估計中位存活時間的偏
誤及均方誤。最後，本文藉一筆資料說明所提模型之應用。

The parametric accelerated failure time(AFT) model for analyzing the right-censored survival data usually assumes Weibull, lognormal or log-logistic distribution for the lifetime variable of interest. When the covariates are presented, the AFT model is often expressed log-linear
form. In this study, however we consider the parametric accelerated failure time model with lifetime which is distributed for a generalized gamma distribution, including, in particular, Weibull, lognormal and gamma distributions. In addition, we use local likelihood estimation to investigate the possible transformation of the covariates that showing in the log-linear model. A simulation study is further implemented for evaluating the bias and mean squares error of the estimated medians for a
variety of AFT models and survival distributions. Finally, we use a data to illustrate the application of the model, under study.

[1]Bagdonavicius, V. and Nikulin, M.(2001). Accelerated Life models Modeling and Statistical Analysis. Springer: New York.
[2]Cleveland, W. S.(1979). Robust Locally Weighted Regression and Smoothing Scatterplots, Journal of the American Statistical Association, 74, 828-836.
[3]Cox ,C.,Chu, H., Schneider, M. F. and Munoz, A.(2007). Parametric survival analysis and taxonomy of hazard functions for the generalized gamma distribution, Statistics in Medicine, 26, 4352-4374.
[4]Fleming, T. R. and Harrington, D. P.(1991). Counting Processes and Survival Analysis. Springer: New York.
[5]Friedman, J. H. and Stuetzle, W.(1981). Projection Pursuit Regression, Journal of the American Statistical Association, 76, 817-823.
[6]Friedman, J. H. and Stuetzle, W.(1981). Smoothing of Scatterplots, technical report(Orion 003), Stanford University, Dept. of Statistics.
[7]Mudholkar, G. S., Srivastava, D. K. and Kollia, G. D.(1996). A generalization of the Weibull distribution with application to the analysis of survival data, Journal of the American Statistical Association, 91, 1575-1583.
[8]Tibshirani, R. and Hastie, T.(1987). Local Likelihood Estimation, Journal of the American Statistical Association, 82, 398., 559-567
[9]Sparling, Y. H., Younes, N. and Lachin, J. M.(2006). Parametric survival models for interval-censored data with time-dependent covariates, Biostatistics, 7, 4, 599-614.
[10]Stacy, E. W.(1962). A generalization of the gamma distribution, Annals of Mathematical Statistics, 33, 1187-1192.
[11]Stacy, E. W. and Mihram G. A.(1965). Parameter estimation for a generalized gamma distribution, Technometrics, 7, 349-358.
[12]Wei, L. J.(1992). The accelerated failure time model: a useful alternative to the Cox regression model in survival analysis, Statistics in Medicine, 11, 1871-1879.