Statistical Inference Of AH Model Based On GEE-Copula Clustering Censored Data

In the context of medical research,clustering interval censored data is an important data type.Each type of data has a censored structure and the data is often accompanied by a strong correlation.It is necessary to deal with the correlation between data.In this paper,the correlation between the data is characterized by the Copula function and expressed by the work correlation matrix in the generalized estimation equation,and the GEE-Copula estimation method is obtained,which is used to statistically infer the additive risk model.Using numerical simulation to verify the effectiveness of the method,and at the same time comparing and analyzing with the estimation results obtained by the generalized estimation equation under independent structure,it is found that the estimation results obtained by GEECopula are better than the estimation validity and stability.The result obtained by the generalized estimation equation under the independent structure.The first two chapters of this article mainly explore the research background and significance of the selected topics and briefly describe the basic theories and methods of application.The main contents are chapters three,four and five:Chapter 3 analyzes the clustering failure data through the example of gastric cancer precancerous lesions,and estimates the interest parameters in the Logit regression model by the method of generalized estimation equations..Finally,the QIC criterion is used to select the most suitable work correlation matrix under the data.Chapter 4 explores GEE-Copula estimation of the type I censoring in the clustering interval.Firstly,the data form of the type I censored data in the clustering interval is introduced.The AH model is established.The Copula function is used to describe the relevant structure in the generalized estimation equation.The parameters of the system of equations are solved simultaneously by Newton's iterative algorithm.The simulation shows that the method is effective.Chapter 5 explores GEE-Copula estimation of the type II censoring in the clustering interval.First introduces the data form of the type II censored clustering interval,establishes the AH model,selects the Copula function to describe the working matrix in the generalized estimation equation,constructs the estimation equation of the regression parameter in the AH model and the estimation equation of the connection parameter in the Copula function,and finally The parameters of the system of equations are solved by Newton's iterative algorithm.The simulation shows that the method is effective.