Research Thrust

Rapid Detection of Emerging Patterns

The ability to rapidly detect and identify unanticipated emerging patterns and trends in data can be highly beneficial in many practical situations. For instance, a swift detection of an outbreak of an infectious disease in its early stage may give public health services a chance to react promptly, mitigating social and economical consequences of such an event. Early detection of a pattern of systematic failures of equipment may preempt possible shortages of certain spare parts and help in faster diagnosis of the root cause of the problem. Such benefits can be realized by sifting through streams of relevant data, such as daily sales of particular kinds of medicines, counts of patients reporting to emergency rooms with particular symptoms, or equipment repair and maintenance records.  Appropriately designed analytic algorithms and software are key to success. Auton Lab statistical data mining algorithms have demonstrated the ability to detect such emerging patterns in medical, agricultural and sales data. Empirical results of their evaluation in practical real-world setups typically indicate substantially higher detection power at lower false alarm rates, and one to three orders of magnitude faster computation than demonstrated by previous apporaches. Auton Lab algorithms for rapid detection of emerging patterns currently include WSARE, Fast Spatial Scan, and TipMon. They are being successfully used by public health officials in the US and abroad, by food safety inspectors, and they are being evaluated by practitioners from other domains. 

Papers 

Name	Authors	Actions
A Bayesian scan statistic for spatial cluster detection	Daniel Neill, Andrew Moore, Gregory Cooper	show
A Bayesian spatial scan statistic	Daniel Neill, Andrew Moore, Gregory Cooper	show
A Fast Multi-Resolution Method for Detection of Significant Spatial Disease Clusters	Daniel Neill, Andrew Moore	show
A Fast Multi-Resolution Method for Detection of Significant Spatial Overdensities	Daniel Neill, Andrew Moore	show
A Study into Detection of Bio-Events in Multiple Streams of Surveillance Data	Josep Roure, Artur Dubrawski, Jeff Schneider	show
Bayesian Network Anomaly Pattern Detection for Disease Outbreaks	Weng-Keen Wong, Andrew Moore, Gregory Cooper, Michael Wagner	show
Detecting Anomalous Patterns in Pharmacy Retail Data	Robin Sabhnani, Daniel Neill, Andrew Moore	show
Detecting Significant Multidimensional Spatial Clusters	Daniel Neill, Andrew Moore	show
Efficient Algorithms for Non-Parametric Clustering with Clutter	Weng-Keen Wong, Andrew Moore	show
Efficient Analytics for Effective Monitoring of Biomedical Security	Robin Sabhnani, Daniel B. Neill, Andrew W. Moore, Artur W. Dubrawski, Weng-Keen Wong	show
Monitoring Food Safety by Detecting Patterns in Consumer Complaints	Artur Dubrawski, Kimberly Elenberg, Andrew Moore, Maheshkumar Sabhnani	show
Optimal Reinsertion: A new search operator for accelerated and more accurate Bayesian network structure learning	Andrew Moore, Weng-Keen Wong	show
Rapid Detection of Significant Spatial Clusters	Daniel Neill, Andrew Moore	show
Rule-based Anomaly Pattern Detection for Detecting Disease Outbreaks	Weng-Keen Wong, Andrew Moore, Gregory Cooper, Michael Wagner	show
Summary of Biosurveillance-relevant statistical and data mining technologies	Andrew Moore, Gregory Cooper, Rich Tsui, Michael Wagner	show
What's Strange About Recent Events	Weng-Keen Wong, Andrew Moore, Gregory Cooper, Michael Wagner	show