The thesis is a research on a parameter estimate of contamination data for simple linear regression model.In the first part,we study simple linear regression model x_i~[(o]) = γ+βμ_i + ε, (i = 1,2,…, n) when its respone random variables{x_i~[(o)] } are contaminated by another i.i.d. random variable sequence{ω_i}. We are only observe contaminated data. So,we are give estimations of parameter model and coefficient of contamination.Under suitable condition we proved that the estimations which are given in the thesis are strongly consistent estimators.In the second part,In contamination regression model.we only suppose that error having k-th momment. we set up strongly consistence of contamination coefficient and regression coefficient.
|