The exponentially weighted moving average (EWMA) statistic, originally developed by Roberts (1959), provides a mechanism to integrate both historical and current data to increase estimation accuracy. Motivated by this concept, the current study introduces memory-type exponential ratio and product estimators of the population mean under simple random sampling (SRS) that rely on the EWMA framework. To determine efficiency conditions, the proposed estimators’ bias and mean square error (MSE) analytical formulas are learned and compared to the features of conventional and well-known memory-type estimators. A thorough simulation study along with an empirical analysis based on actual data sets are conducted to further evaluate their practical applicability, and both demonstrate the superiority of the suggested methodology.
