Modeling Method Research Based On Longitudinal And Survival Data And Its Application In Liver Cirrhosis Data

In the production practice and scientific research,we often meet with complex data of repeated measure or related to the survival time.Longitudinal data is a kind of repeated measure data and survival data is a kind of data related to the survival time.Studying the survival and longitudinal data separately,we can easily understand,but the correlation between them and biased estimation produced by analysis are often ignored.Therefore,some scholars proposed joint model to combine the two kinds of data model.In the process of research,the event time is often censored causing by the limitations of time,economy,measuring tools,and we can't obtain the full longitudinal data.The noteworthy feature of survival data is censored,therefore,the situation of data censored should be taken into account in the process of building the joint model.Using the joint model of longitudinal and survival data to analyze has a certain practical value.In this paper,the mainly work is described as follows:In the first chapter,we introduces the characteristics of survival data and its basic function,the commonly used model,the definition and characteristics of longitudinal data and the research review of model method.We also illustrate the research background and significance of this study,and explain the data used in this paper.In the second chapter,we introduces the Cox proportional hazards model of survival data and linear mixed effects model of longitudinal data,and their respective model parameter estimation method,preparing for the joint model in the next chapter.In the third chapter,we build the standard joint model of longitudinal and survival data based on survival data model and longitudinal data model introduced in the second chapter,and extend the standard joint model,and present the model selection criterion,and finally give the maximum likelihood estimation method of the model parameters.In the fourth chapter,we build the generalized joint model of longitudinal and survival data based on the third chapter,and extend the generalized joint model,and give the model selection criterion,and finally give the bayesian estimation method of the model parameters.In the fifth chapter,we use the PBC data to analysis separately and jointly.The results show that joint analysis is better than single analysis,and the results of bayesian estimation is better than the maximum likelihood estimation.