Regression Analysis Of Clustered Data, Under The Additive Transformation Models

Since survival analysis started to develop from the last century,with the progress of its theory and application,it has become a very important means for the statistical analysis of incomplete data.This paper first introduces the research background and literature reviews of survival analysis.After providing some preliminary knowledge of types of survival data,basic functions established for survival analysis and some semi-parametric models,the focus has been on regression analysis of clustered data arising from a class of additive transformation models "which are generalized from additive hazards models.In particular,we consider the situation where the correlated failure times of interest may be related to cluster sizes.That is,the cluster sizes are informative.To address the problem,we present an existing method for clustered right-censored data and provide estimation procedures based on the weighted estimating equation(WEE-based)method for clustered current status data.The asymptotic properties of the WEE-based estimates are established.To assess the finite sample performance of the proposed WEE-based method,extensive simulation studies are conducted,indicating that the method works well.Respective applications to two sets of real data arising from a tumorigenicity study and a skin cancer study are also provided.