Font Size: a A A

Bayesian Analysis Of Mixed Model Of Semicontinuous Longitudinal Data Based On MCMC Algorithm

Posted on:2021-01-19Degree:MasterType:Thesis
Country:ChinaCandidate:D Q LuFull Text:PDF
GTID:2480306230480124Subject:Applied Statistics
Abstract/Summary:PDF Full Text Request
Semicontinuous longitudinal data is common in many fields,such as epidemiology,economics,sociology,etc.,and the mixed-effect joint model of semicontinuous longitudinal data is getting more and more attention in the academic circles.This joint model allows the random effects in the model to be correlated flexibly.Bayesian statistical inference has been well developed and applied in many fields,but it has not been applied to the above models.Therefore,in this paper we applies the theory and method of Bayesian statistical inference to the joint model of semicontinuous longitudinal data.In this paper,under the framework of Bayesian theory,we mainly conduct in-depth research on the construction of mixed-effect joint model of semicontinuous longitudinal data,parameter estimation and its application in medical problems.The model is constructed according to the characteristics of semicontinuous longitudinal data distribution.The constructed model consists of two mixed effect models: one model is a logistic mixedeffect model,and the other model is a linear mixed-effect model.These two models are related by random effects.At the same time,based on the likelihood function of the model,we combined with the prior information,the Bayesian posterior inference and MH algorithm for the mixed-effect joint model of semicontinuous longitudinal data are given.Then,the relevant posterior distribution is converted into the form of a matrix to improve the calculation efficiency.For the posterior distribution of parameters,we conducted a numerical simulation using a hybrid algorithm of Gibbs sampling and MH sampling,and then gave Bayesian estimates of the parameters in the model.By comparing with the real value,the estimated effect of the parameter estimates is analyzed,and the estimated efficiency under different prior information is compared.In this paper,the MH algorithm and Gibbs algorithm are used in combination to study the Bayesian estimation method of the mixed-effect model of semi-continuous longitudinal data.At the same time,the method is applied to the pathological study of prostate cancer,and the influencing factors of the treatment effect of prostate cancer are analyzed.
Keywords/Search Tags:Semi-continuous longitudinal data, Bayesian estimation, MetropolisHastings algorithm, Gibbs algorithm, Mixed effect model
PDF Full Text Request
Related items