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:Penalized optimal design for dose finding - Pronza
to\, L (CNRS)
DTSTART;TZID=Europe/London:20110816T100000
DTEND;TZID=Europe/London:20110816T103000
UID:TALK32380AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/32380
DESCRIPTION:We consider optimal design under a cost constraint
\, where a scalar coefficient L sets the compromis
e between information and cost. For suitable cost
functions\, one can force the support points of an
optimal design measure to concentrate around poin
ts of minimum cost by increasing the value of L\,
which can be considered as a tuning parameter that
specifies the importance given to the cost constr
aint. \n\nAn example of adaptive design in a dose-
finding problem with a bivariate binary model will
be presented. As usual in nonlinear situations\,
the optimal design for any arbitrary choice of L d
epends on the unknown value of the model parameter
s. The construction of this optimal design can be
made adaptive\, by using a steepest-ascent algorit
hm where the current estimated value of the parame
ters (by Maximum Likelihood) is substituted for th
eir unknown value. Then\, taking benefit of the fa
ct that the design space (the set of available dos
es) is finite\, one can prove the strong consisten
cy and asymptotic normality of the ML estimator wh
en L is kept constant. Since the cost is reduced w
hen L is increased\, it is tempting to let L incre
ase with the number of observations (patients enro
led in the trial). The strong consistency of the M
L estimator is then preserved when L increases slo
wly enough.\n\n\n
LOCATION:Seminar Room 1\, Newton Institute
CONTACT:Mustapha Amrani
END:VEVENT
END:VCALENDAR