随机删失下指标系数的半参数估计

SEMIPARAMETRIC ESTMATION OF INDEX COEFFICIENT UNDER RANKOM CENSORSHIP

没有模型Ｙ＝ｇ（Ｘ′β）＋ｅ，其中Ｘ为ｐ－维解释变量，β为ｐ维未知参数ｇ（·）为一元未知函数，ｅ为随机误差，Ｅｅ＝０．本文在随机右删失数据，之下构造了指标系数β的方向估计，证明 具有 相合性及渐近正态性．

Det Y=g(Xβ)+e,where X(∈R)is an explanatory variable,β(∈R,)an unknownparmeter,g(·)an unknown univariate function,e the random error with Ee =0.In this paper,theestimator δ_n for the direction of index coefficient βis constructed under the random right censoreddata Zi=min(Y_i,W_i)=I(Y≤W_i)(i=1,2,…,n)and it is shown that is-consistent and asymptotically normal.