Empirical Likelihood Confidence Regions of Parameters in Nonlinear EV Models under Missing Data
缺失数据下非线性EV模型参数的经验似然置信域

The missing response problem in the nonlinear EV(error-in-variables) models is considered, where the explanatory variate X is erroneously measured. With the help of val-idation data, two empirical log-likelihood ratio statistics for the unknown parameters in the model are proposed. It is proved that the proposed statistics are asymptotically chi-square distribution under some mild conditions, and hence can be used to constructing the confidence regions of the parameters.