Optimal Parallel Sequential Change Detection Under Generalized Performance Measures

This paper considers the detection of change points in parallel data streams, a problem widely encountered when analyzing large-scale real-time streaming data. Each stream may have its own point, at which has distributional change. With sequentially observed data, decision maker needs to declare whether changes already occurred streams each time point.Once is declared changed, it deactivated permanently so that future will no longer be collected. compound sense want optimize certain performance metrics concern all as whole. Thus, decisions are not independent for different streams. Our contribution three-fold. First, we propose general framework includes ones considered existing works special cases and introduces new connect closely with single-stream sequential hypothesis testing. Second, data-driven procedures developed under this framework. Finally, optimality results established proposed procedures. The methods theory evaluated by simulation studies case study.