Estimation In Partially Linear Varying-coefficient Errors-in-variable Model With Restricted Condition Under Missing Responses

As parameter model to the nonparametric model to the semiparametric model ofdevelopment all the way, in recent years,semiparametric partially linear varying-coefficient model is a kind of new model of dealing with high-dimensional data inregression analysis. Linear model, partially linear model and varying-coefficient modelare all degradation forms of semiparametric partially linear varying-coefficient model.Compared with the linear model, partially linear varying-coefficient model has reducedthe modeling deviation substantially and avoided the curse of dimensionality. The abovecan be considered as a wider range of applications of the model, not only can be used toanalyse independent data, but also longitudinal data and time series data and othercomplex data effectively.Measurement errors often are inevitable in actual problems. if ignoring themeasurement errors and directly considering semiparametric partially linear varying-coefficient model, the obtained estimation is often biased. Therefore, the introduction ofsemiparametric partially linear varying-coefficient errors-in-variables model is of greattheoretical significance and practical value.Missing datas and restricted condition are two types of problems in the actualoperation. Due to various of subjective factors and objective factors it often encountersthe missing datas phenomenon in obtaining data, so how to deal with the missing datasnaturely becomes a valuable problem to research. And during the process of statisticalanalysis, the information of parameters or non-parameters is usually not unknown,during the process of the past experiences, it can often obtain some prior informationwhich be called constraint conditions of parameters and non-parameters.This article studies partially linear varying-coefficient errors-in-variable modelwith restricted condition under missing responses. According to thought of Ahmad Iabout semiparametric partially linear varying-coefficient errors-in-variables model inthe case that the random error is conditionally heteroskedastic, we obtation theestimation of parameters and non-parameters by general series estimation method, anddiscusse the consistency and asymptotic normality of parametric estimator and the rateof convergence in nonparametric part. Moreover the estimates of the proposed method is verified by the numerical simulation results with R software under limited samples byMonte Carlo.