Displaying results 1 - 4 of 4
-
Anomaly Pattern Detection for Biosurveillance
Content Type: Abstract
We propose a new method for detecting patterns of disease cases that correspond to emerging outbreaks. Our Anomaly Pattern Detector (APD) first uses a "local anomaly detector" to identify individually anomalous records and then searches over subsets… read more… visiting nearby hospitals with similar symp- toms), and may not be evident by examination of any single record … from Alleg- heny County, PA. Figure 1 plots the detection preci- sion, i.e. the proportion of detected anomalies that … The Bayesian aerosol release detector. Stat. Med., 2007, 26: 5225-5252. Further Information: Daniel B. Neill, … -
Searching for Complex Patterns Using Disjunctive Anomaly Detection
Content Type: Abstract
Modern biosurveillance data contains thousands of unique time series defined across various categorical dimensions (zipcode, age groups, hospitals). Many algorithms are overly specific (tracking each time series independently would often miss early… read more… of outbreaks), or too general (detections at state level may lack specificity reflective of the actual process at … dimension. An example anomalous cluster detectable by DAD may identify zipcode = {z1 or z2 or z3 or z5} and age_group … of outbreaks), or too general (detections at state level may lack specificity reflective of the actual process at … -
Detection of multiple overlapping anomalous clusters in categorical data
Content Type: Abstract
Syndromic surveillance typically involves collecting time-stamped transactional data, such as patient triage or examination records or pharmacy sales. Such records usually span multiple categorical features, such as location, age… read more… Weekly Epidemiological Reports3. The data stores patient visits spanning 26 regions and 9 diseases reported over 2.5 years. We … -
Learning Specific Detectors of Adverse Events in Multivariate Time Series
Content Type: Abstract
This paper describes how powerful detectors of adverse events manifested in multivariate series of bio-surveillance data can be learned using only a few labeled instances of such events.… by being able to raise an alert even if none of the com- ponent signals is critical, but if some of them are … powerful specific detectors of adverse events in mul- tivariate time series data, even if the number of …