Change Point Estimation And Application Of Censored Quantile Regression Model

This paper mainly studies the change point estimation and application of censored quantile regression model.The study aims to estimate the number,location and regression coefficient of the change points in the model.The validity and robustness of the estimation are verified by numerical simulation.The proposed method is applied to analyze real data.In the first chapter,the research background and status of quantile regression model,censored quantile regression model and change point are introduced.In the second chapter,the left censored quantile regression model with single change point is studied by the grid search method.The change point and regression coefficient are estimated,and the large sample properties of the estimator are obtained.The effectiveness and robustness of the estimation are verified by numerical simulation.The proposed method is applied to empirical analysis of household financial assets data,and giving a reasonable explanation of the results.In the third chapter,we study the change point estimation in right censored quantile regression model by linearization technique.Linearization technique is easy to understand,simple and effictive,and can be extended to multiple change points.Firstly,the linearization technique is used to estimate the single change point and regression coefficients.Secondly,we estimate change point in the right censored quantile regression model.Using BIC information criterion to determine the number of unknown change points,estimate change points and coefficient estimation via linearization technique.Finally,simulation results show that the proposed estimator has desirable finite sample performance.In the fourth chapter,this paper analyzes two actual datasets using the aforementioned model,one is an empirical study of health status in China,the other is analysis on the actual problem of BMI(Body Mass Index)and SBP(Systolic Blood Pressure).Firstly,the number of unknown change points are determined.Then,the estimates of change points and coefficients are obtained.Finally,the empirical analysis shows that there are multiple change points at different quantile.In the last chapter,the main conclusions of this paper are summarized,and the future work to be continued is also analyzed.