Monitoring Complex Segmented Streams of Data Using Bootstrap Control Charts

Monitoring complex streams of data has been considered. The application of well-known statistical process control (SPC) methods, such as Shewhart or CUSUM control charts, may be for such processes questionable. In the paper, we consider processes consisting of segments and subsegments. The data from subsegments belonging to respective segments are aggregated using probabilistic and possibilistic methodologies. We use a bootstrap control chart for monitoring such aggregated streams of data. Using computer simulations, we show that the monitoring of aggregated data is efficient and interpretable. The motivation for this research comes from a real-life problem—monitoring bipolar disorder psychiatric patients using their measured voice characteristics.