Statistical Analysis For Several Special Types Of Varying-coefficient Models

In this thesis, we have mainly studied several special varying- coefficient models , including the functional-coefficient partial linear models, the varying-coefficient models with AR(1) error , the varying- coefficient models withφ-mixing error .Full thesis has five chapters:The first chapter, we have introduced the research present situation of varying-coefficient models,the common nonparametric estimated method,the error is the autocorrelation model and the research present situation.The second chapter, we have studied the local linear estimate for functional-coefficient partial linear models under the improvement condition, and obtained the convergent rate and asymptotic normality for estimate of coefficient function under the constant funciont estimated by B-spline method finally.The third chapter, we discussed the varying-coefficient models under error is linear AR(1) and the error is nonparameter AR(1) respectively,We gave the local linear estimate for coefficient function, and the convergent rate for the estimate is proved.The fourth chapter, we discussed estimate for varying-coefficient models withφ-mixing error and proved the estimate is consistency.The fifth chapter,we summary this thesis.