BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Modelling Astrophysics Data for Discovery\, Classification and Pre
 cise Measurements - David Hogg (Centre for Cosmology &amp\; Particle Physi
 cs\, New York University)
DTSTART:20110301T163000Z
DTEND:20110301T173000Z
UID:TALK29613@talks.cam.ac.uk
CONTACT:David Titterington
DESCRIPTION:In applications as varied as the measurement of stellar proper
  motions\, the determination of the Milky Way mass with maser kinematics\,
  and the selection of quasar targets for SDSS-III BOSS\, precise - and mor
 e important\, accurate - data analysis requires a model that generates the
  data. \nA generative model produces a probability distribution function i
 n the space of the noisy data\, after convolution by observational uncerta
 inty distribution functions.  I show that proper modelling of the data-gen
 erating process performs better than other data analysis and\nclassificati
 on methods\, in scientific applications in \nwhich measurements come with 
 relatively reliable uncertainty estimates.  I make also some comments on t
 he\ntheoretical basis for and ideal outputs from any \nprincipled program 
 of data analysis.  These results have implications for almost all ongoing 
 and future \nastrophysics projects.
LOCATION:Ryle Seminar Room\, Cavendish Laboratory
END:VEVENT
END:VCALENDAR