Researchs on Regression Model with Dependent Error
|School||Hefei University of Technology|
|Keywords||φ-mixing (ρ ~ ) - Mixed (φ ~ ) - Mixed PQD sequence Kernel estimation Consistency|
The main problem is a regression model is the inference of the regression parameter β , such as the estimated amount of consistency, a diagnosis is based on the the β estimated amount of error distribution . Which many classical results are established in a separate case , the premise of independence does not hold , but on many occasions . Therefore, in the dependent sample , scholars have done a lot of work . The literature existing results further analysis , and get some good results . The details are as follows : a study based on the asymptotic nature of the Linear model error distribution dependent error , constitute the second chapter . On the basis of the second chapter , the third chapter , we further study based on M-estimators under an estimate of the error distribution of the linear model , the estimated amount Consistency weaker conditions , promotion of the existing literature related results. Three in the fourth chapter , we study the problems of nonlinear regression model regression parameter estimates based on samples positive quadrant dependent (PQD) , Strong Consistency of the estimated amount of results .