Quantile regression and its empirical likelihood with missing response at random (Q725686)

The authors consider a linear quantile regression model with missing response data, under the assumption of missing at random. For the missing data, they employ the inverse probability weighted method, rather than imputation. On the basis of an estimation equation on quantile regression they establish, a standard quantile regression estimator for the covariate vector is given, as well as its empirical likelihood estimator. Under suitable assumptions, the asymptotic normality of these estimators is established. Furthermore, the authors construct the appropriate empirical likelihood ratio statistic, and show that its asymptotic distribution is chi-square.