New modification of ranked set sampling for estimating population mean

Abstract

A key problem in statistics and machine learning is drawing samples from a population. The drawn sample is mainly used to make inferences about the factors and characteristics of the population in order to make high-accuracy estimations. This study developed a new sampling method called Except Extreme Ranked Set Sampling (EERSS) for estimating the population mean. EERSS estimator is compared with its counterparts by using Simple Random Sampling (SRS), Ranked Set Sampling (RSS), Median Ranked Set Sampling (MRSS), and Moving Extreme Ranked Set Sampling (MERSS). Simulation study and real data sets are used for illustrations and comparisons. It shows that the EERSS estimator is an unbiased estimator of the population mean for symmetric distributions and it is the most efficient among all estimators.