BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//talks.cam.ac.uk//v3//EN
BEGIN:VTIMEZONE
TZID:Europe/London
BEGIN:DAYLIGHT
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
TZNAME:BST
DTSTART:19700329T010000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
TZNAME:GMT
DTSTART:19701025T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
CATEGORIES:Isaac Newton Institute Seminar Series
SUMMARY:Bayesian experimental design for stochastic dynami
cal models - Gibson\, GJ (Heriot-Watt University)
DTSTART;TZID=Europe/London:20110721T113000
DTEND;TZID=Europe/London:20110721T123000
UID:TALK32118AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/32118
DESCRIPTION:Advances in Bayesian computational methods have me
ant that it is now possible to fit a broad range o
f stochastic\, non-linear dynamical models (includ
ing spatio-temporal formulations) within a rigorou
s statistical framework. In epidemiology these me
thods have proved particularly valuable for produc
ing insights into transmission dynamics on histori
cal epidemics and for assessing potential control
strategies. On the other hand\, there has been les
s attention paid to the question how future data s
hould be collected most efficiently for the purpos
e of analysis with these models. This talk will de
scribe how the Bayesian approach to experimental d
esign can be applied with standard epidemic models
in order to identify the most efficient manner fo
r collecting data to provide information on key ra
te parameters. Central to the approach is the repr
esentation of the design as a 'parameter' in an ex
tended parameter space with the optimal design app
earing as the marginal mode for an appropriately s
pecified joint distribution. We will also describe
how approximations\, derived using moment-closure
techniques\, can be applied in order to make trac
table the computational of likelihood functions wh
ich\, given the partial nature of the data\, would
be prohibitively complex using methods such as da
ta augmentation. The talk will illustrate the idea
s in the context of designing microcosm experiment
s to study the spread of fungal pathogens in agric
ultural crops\, where the design problem relates t
o the particular choice of sampling times used. We
will examine the use of utility functions based e
ntirely on information measures that quantify the
difference between prior and posterior parameter d
istributions\, and also discuss how economic facto
rs can be incorporated in the construction of util
ities for this class of problems. The talk will d
emonstrate how\, if sampling times are appropriate
ly selected\, it may be possible to reduce drastic
ally the amount of sampling required in comparison
to designs currently used\, without compromising
the information gained on key parameters. Some cha
llenges and opportunities for future research on d
esign with stochastic epidemic models will also be
discussed.\n
LOCATION:Seminar Room 1\, Newton Institute
CONTACT:Mustapha Amrani
END:VEVENT
END:VCALENDAR