Researchs On Regression Model With Dependent Error

For the regression model,one of the most important questions of this model is the inference of the regression parameterβ, such as the consistence of the estimator, and the diagnosis of error distribution. In the regression analysis,many classical results are based on independent assumption.However,this assumption does not hold in many occasions.Therefore, scholars make many efforts under dependent samples. In this dissertation, we make further analysis of presence conclusions in existing literature, and obtain some good results.The details are given as follows:1. This dissertation researches asymptotic properties of error distribution on the basis of dependent error, which constitutes of the second section.2. In the third section, we further research the estimation problem of error distribution in linear model based on M estimation,get the consistence of the estimator under weaker conditions and promote some related results of existing documents.3. In the forth chapter, we research the estimation problem of parameter in nonlinear regression model on the basis of positive quadrant dependent samples,construct estimator and get the result of strong consistence of estimator.