Estimation Method Of AR(p) Model With Data Missed

In this paper,we deals with the parameter estimation problem for AR(p) model with data missing.In the literature,EM algorithm or MCMC method is used to solve the estimation problem for one data or two successive data missing.But it is difficult to deals with the estimation problem for several successive data missing,because computational complexity.In this paper,we use Regression Forecast Method and least-squares estimation to deal with the parameter estimation problem with one data or two successive data missing.Then we consider with the parameter estimation problem with several successive data missing and present a practical method.By the method,we have obtained the estimated values of parameter. Then we discuss how to solve the general situation about the problem by application examples. Finally,numerical simulation about the parameter estimation problem with one data or two successive data missing is done.There are three parts in this paper:Section 1,Least-squares estimation problem for AR(p) model parameter;Section 2,Parameter estimation methods with data missing;Section 3,Sample application and numerical simulation.