Related Experiment Videos
WSARE: What's Strange About Recent Events?
Weng-Keen Wong1, Andrew Moore, Gregory Cooper
1Department of Computer Science, Carnegie Mellon University, Pittsburgh, PA 15213, USA. wkw@cs.cmu.edu
Summary
This study introduces a new algorithm for early disease outbreak detection using emergency department data. It effectively identifies anomalous illness patterns, improving detection timeliness over traditional methods.
Area of Science:
- Public Health
- Computer Science
- Epidemiology
Background:
- Traditional anomaly detection methods identify rare individual events, not group illness patterns.
- Existing techniques are insufficient for early detection of disease outbreaks.
Purpose of the Study:
- To develop and evaluate a novel algorithm for early disease outbreak detection.
- To identify anomalous patterns of illness in emergency department data.
Main Methods:
- Proposed an anomaly detection algorithm characterizing patterns with rules.
- Evaluated rule significance using Fisher exact and randomization tests.
- Compared the new algorithm against a standard detection method using simulated epidemic data.
Main Results:
- The proposed algorithm demonstrated significantly improved detection times.
- A slightly higher false positive rate was observed compared to the standard algorithm.
- The algorithm effectively detected anomalous group illness patterns.
Conclusions:
- The new algorithm offers enhanced timeliness for early disease outbreak detection.
- The approach shows promise for improving public health surveillance systems.
- Further refinement may address the false positive rate.