Real Time Monitoring for Abnormal Events with an Application to Influenza Outbreaks

Wednesday 23 October 2013, 1.00PM to 2.00pm

Speaker(s): Brendan McCabe, University of Liverpool

Abstract:

We monitor dependent sequences for abnormal activity in real time.  The "deviation" metric is based on p-values for the system in control and no model for the type of abnormality is required. After an “Alert” has been declared we can continue to monitor to declare an "All Clear".   Possible applications include monitoring for  fraudulent transactions;   financial/ housing bubbles etc; disease outbreaks/epidemics and so on. An extended application is given to detecting the H1N1 flu virus is given.

Location: ARRC Auditorium (A/RC/014)