On enhanced monitoring of paired data using memory control chart: application in plant sciences

Monitoring small shifts in paired observations is essential in disciplines such as plant sciences, where measurements like sepal and petal dimensions are naturally correlated. Conventional control charts often assume independence, which can lead to misleading results in such contexts. This study proposes a bivariate exponentially weighted moving average (EWMA) control chart tailored for correlated paired data. The chart integrates the inherent correlation structure, thereby enhancing sensitivity to subtle process shifts compared to the traditional Shewhart chart. A comprehensive performance evaluation is conducted through Monte Carlo simulations under three shift scenarios, using individual and overall performance measures. The results demonstrate superior detection ability of the proposed chart across varying parameter settings. An application to the classical Iris dataset further illustrates its practical effectiveness in capturing small but meaningful variations in biological processes. The findings highlight the usefulness of the proposed chart for monitoring paired data in biological, environmental, and industrial studies, with potential extensions to broader multivariate settings.