在响应变量随机缺失耐,运用借补拟似然方法估计广义半参数模型的未知参数和非参数回归函数,验证了所给出的估计具有渐近正态性,并进行了模拟研究.通过模拟结果可知,本文通过回归方法借补缺失值,再利用全部数据进行估计的方法比完整个体分析的方法要好.
We propose a local imputation quasi-likelihood estimator for generalized semiparametric models when the responses are missing at random. Furthermore, we get the asymptotically normal of the estimator with given the asymptotic biases and variances of the estimation. Through simulation we can see that the imputation estimation is better than complete case estimation.