误差为鞅差序列的半参数回归模型估计的相合性

On Consistencey of a Semiparametric Regression Estimate under Martingale Difference Sequence

设半参数回归模型Ｙ（ｎ）ｉ＝β·ｘ（ｎ）ｉ＋ｇ（ｔ（ｎ）ｉ）＋Ｅ（ｎ）ｉ，ｉ＝１，２，…，ｎ，本文由最小二乘法和一般加权方法定义的β、ｇ（ｔ）的估计量βｎ，ｇｎ（ｔ），在误差为鞅差序列下获得了βｎ，ｇｎ（ｔ）的ｒ（≥２）阶平均相合性．

Considered Semiparametric regression model Y (n) i=β·X (n) i+g(t (n) i)+ε (n) i , We use the least squares and usual weighted method to define the estimates n and n for β and g and obtain its r(≥2) th mean consistencey under martingale difference sequence.