Font Size: a A A

Research On Parameter Regression Model Of Interval Censored Data In Proportional Hazard Model

Posted on:2018-12-22Degree:MasterType:Thesis
Country:ChinaCandidate:S Y LinFull Text:PDF
GTID:2347330536459562Subject:Statistics
Abstract/Summary:PDF Full Text Request
At present,the most commonly used methods in the survival analysis of COX proportional hazards model,this model has been widely applied in the study of chronic epidemiology,and based on the parameter proportion hazard regression model of various type distribution in biology,medicine,engineering,science and sociology,psychology,economics,actuarial science and reliability and other fields have an important tool role.In this paper,we focus on two types of interval censored data(interval censored case I and interval censored case II),and establish the generalized exponential COX proportional hazard model.We choose the generalized exponential distribution,because the distribution can do better than the Weibull distribution and the gamma distribution in many aspects,the hazard function is very flexible and can be applied to many types of modeling and analysis of risk rate data.Of course,for interval censored time data,it also has a better and more flexible analysis effect,and has important applications in many fields such as life test and reliability study.Focus on innovative research from two aspects.Firstly,the interval censored data type II,based on a generalized exponential proportional hazard regression model,due to the complexity of the data,the estimation results obvious method cannot give maximum likelihood estimate model,parameter estimation algorithm using Newton-Rapson proportional hazard model,through the model of a large number of setting various situations,to verify the validity of the model and estimation of the proposed method.The model was applied to the analysis of the clinical trials of AIDS treatment among 31 AIDS patients.Secondly,the interval censored I data to establish a generalized exponential distribution proportional hazards regression model,because of the complexity of the data and the likelihood function,so try to use prior information with Bayesian estimation method to estimate the parameters,given the hierarchical posterior distribution showed no expression,application of various algorithms MCMC sampling,Bayesian estimation of the parameters are got.A large number of simulation experiments are designed to verify the effectiveness of the proposed model and algorithm.The model and the Bias estimation algorithm were applied to the actual data and were analyzed in the experimental data of lung tumors in 144 male RFM mice.
Keywords/Search Tags:Proportional hazards model, Interval censored, Bayesian estimation, Maximum likelihood Estimation, MCMC, Newton-Rapson algorithm
PDF Full Text Request
Related items