ABSTRACT
Control chart techniques define and detect outliers and anomalies on a statistical basis. This chapter describes univariate control charts that monitor one variable for anomaly detection. Outliers and anomalies are data points that deviate largely from the norm where the majority of data points follow. The chapter provides a list of software packages that support univariate control charts. Shewhart control charts include variable control charts, each of which monitors a variable with numeric values, and attribute control charts, each of which monitors an attribute summarizing categorical values. Shewhart control charts are sensitive to the assumption that the variable of interest follows a normal distribution. Unlike Shewhart control charts, Shewhart control charts, cumulative sum (CUSUM) control charts and exponentially weighted moving average (EWMA) control charts are effective at detecting anomalies of not only large shifts but also small shifts since CUSUM control charts and EWMA control charts take into account the effects of multiple data observations.
