一类半参数回归模型讨厌参数的相合估计

Consistent Estimator of the Nuisance Parameter in a Semiparametric Regression Model

讨论一类半参数回归模型y=xβ+g(t(1)+t(2)α)+e,其中g(·)未知,α、β为实参数,e为随机误差 设有样本{(ti,xi,yi)}ni)′,构造了g(·)、β和讨厌参数α的估计,并证明了其相合性 i,t(2)i=1,ti=(t(1)

The paper discusses the semiparametric regression medel y=xβ+g(t(1)+t(2)α)=e,where g is an unknown function, α、β are unknown real parametrics, e is a random error with mean 0 and variance σ2. Let (ti,xi,yi),i=1,2,…,n be i.i.d. samples, where ti=(t(1)i,t(2)i)′. Based on the samples proposed the estimators of g、β and the nuisance parametor α, the consistency is obtained.