Research Grant: Using ARIMA Intervention Modeling to Detect Special Causes in Autoregressive Control

To overcome limitations of traditional Shewhart control charts, new control chart methods have been proposed that use ARIMA models to charaterize the time series of the process. However, control limits are typically estimated from global measures of dispersion. If special causes are present in the data used to fit the process control model, the control model parameters will not be accurately estimated. This research empirically tests the use of intervention analysis to identify special causes in autoregressive process data.

Scholars @ ECU