Combining a Moving Average with a Triple EWMA Chart to Improve Detection Performance

This article aims to introduce the novel mixed triple exponentially weighted moving average-moving average (MTEM) chart to accurately detect position changes for both symmetric and non-symmetric distributions. The MTEM chart constructs a moving average (MA) structure to filter out fluctuations in the raw data and then applies triple exponential weighting to improve the ability to identify minor shifts. The average run length (ARL) and median run length (MRL), which are run length profiles derived from the Monte Carlo simulation (MC) strategy, were used to compare the performance of the suggested chart with that of MA, EWMA, TEWMA, and mixed moving average-triple exponentially weighted moving average (MMTE) charts. In addition, the expected average run length (EARL) and expected median run length (EMRL) were also used to rate the overall results. Results of the study indicate that the MTEM chart surpasses competitor charts in detecting minor to moderate changes. The MMTE chart responds slightly slower than the proposed chart. Due to its smoothed and re-averaged structure, it may lose significant information. The MA chart worked better for greater shifts. Furthermore, the MTEM chart competency was applied to two real-world datasets, confirming its practicality.