Jackknife empirical likelihood inferences for the population mean with ranked set samples