Additive distortion measurement errors regression models with exponential calibration

In this paper, we used the newly proposed exponential calibration for additive distortion measurement errors models, where neither response variable nor covariates can be directly observed but are distorted in fashions by an confounding variable. By using calibrated variables, three estimators of parameters and empirical likelihood-based confidence intervals proposed, studied asymptotic properties estimators. For hypothesis testing model checking, adaptive Neyman test statistic restricted is proposed. Simulation studies demonstrate performance statistic. A real example analysed to illustrate its practical usage.