Variable Selection Of Joint Model For Longitudinal Data And Survival Data

In clinical medical research,longitudinal data and survival data often appear simultaneously,and longitudinal data observation processes are always associated with survival events.If the longitudinal data and the survival data are modeled separately,that may lose a lot of information and result in a large estimation error of the model.In order to reduce the error of the model,it is usually necessary to establish a joint model for analyzing data containing both longitudinal processes and survival results.This paper mainly focus on the joint model of linear mixed effect model and additive risk model as submodel,and the variable selection problem of the joint model.For longitudinal data,we establish a linear mixed effect model;for survival data,we establish an addable risk model.Firstly,the two-stage estimation method is used to estimate the regression coefficient of the joint model.Then,the Lasso penalty function and the SCAD penalty function are added to the submodel of the joint model to select and estimate the variables.The comparison with the additive risk model and the joint model with the proportional risk model as the survival sub-model shows the joint model with the additive risk model as the survival sub-model has better fitting goodness.The content of this paper is arranged as follows: The first chapter mainly introduces the research background,the research status of the joint model of longitudinal data and survival data and its variable selection,and show the structure of this paper.The second chapter introduces some basic knowledge needed in this paper.In the third chapter,we propose the joint model of linear mixed effect model and additive risk model and describe the sub-models and parameter estimation methods in detail,and the parameters of cirrhosis clinical data are estimated.The fourth chapter introduces the penalty estimation method of the joint model,and carries out punish estimation and model evaluation on the clinical data of liver cirrhosis.The fifth chapter is the conclusion and prospect.