PEDS-VM: A Variation Management Prototype for Pattern Evolving Data Streams

Widely applied in many domains, data stream processing has attracted more and more attention in database and sensor network communities. In this demo, we present our system PEDS-VM, a real-time surveillance system for managing pattern variation over evolving medical streams. PEDS-VM utilizes an effective strategy and a state-based window framework to extract the evolving patterns. After extraction, PEDS-VM uses a novel storage structure called PGG to efficiently record these incremental patterns. Moreover, several important application scenarios of PEDS-VM are also discussed in this demonstration.