Achievement

Discovering Flow Anomalies for Environmental Forensics:

Flow anomaly discovery identifies time intervals, where readings from adjacent sensors monitoring a stream have significant discrepancies. It is important due to applications in locating previously unknown contamination sources and cleansing mechanisms. Challenges arise from large data volume. Our interdisciplinary team with Computer and Environmental Scientists developed a novel method, namely, the Smart Window Enumeration and Evaluation of persistence-Thresholds (SWEET), using ideas of a smart counter and efficient pruning. Experimental evaluations showed that it significantly improved response time relative to traditional methods making it easier for Environmental Scientists to interactively identify flow anomalies to locate potential pollution sources. Results were accepted for publication in a highly selective forum, namely, the IEEE Intl. Conference on Data Mining, 2008.