Estimation Of Varying Coefficient Transformation Model With Missing Data

Missing data is a common data type in many fields such as medicine,biology,economics finance ec.In this paper,we mainly study the estimation problem of vary-ing coefficient transformation model with missing data.The estimation procedure was divided into two parts,the first part is to study the functional coefficient estimation in the varying coefficient transformation model.First,the B-spline approximation tech-nique is used to approximate the unknown indicate function.Then the varying coef-ficient transformation model becomes a simple linear transformation model,based on the theory of martingale,The estimating equation method is established to conduct the estimation of the functional coefficients in the model.The simulation illustrate the ra-tionality of the method.The second part is to study the model estimation with parital missing covariates.We study the the functional coefficient estimation problem in the varying coefficient transformation model.In this part,the inverse probability weighted method is used to establish the inverse probability weighted estimation equation to es-timate the functional coefficients.Many simulation studies were given to illustrate the performance of our proposed method with finite sample size.In the end of this chapter,we compare the performance with and without missing data and prove the goodness of our estimation.