Data Mining

16

Univariate Control Charts

Outliers and anomalies are data points that deviate largely from the norm where the majority of data points follow. Outliers and anomalies may be caused by a fault of a manufacturing machine and thus an out of control manufacturing process, an attack whose behavior differs largely from normal use activities on computer and network systems, and so on. Detecting outliers and anomalies are important in many fields. For example, detecting an out of control manufacturing process quickly is important for reducing manufacturing costs by not producing defective parts. An early detection of a cyber attack is crucial to protect computer and network systems from being compromised.

Control chart techniques define and detect ...