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:Asymptotic Inference for Eigenstructure of Large C
ovariance Matrices - Jana Jankova (University of C
ambridge)
DTSTART;TZID=Europe/London:20180625T144500
DTEND;TZID=Europe/London:20180625T153000
UID:TALK107368AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/107368
DESCRIPTION:A vast number of methods have been proposed in lit
erature for point estimation of eigenstructure of
covariance matrices in high-dimensional settings.
In this work\, we study uncertainty quantificatio
n and propose methodology for inference and hypoth
esis testing for individual loadings of the covari
ance matrix. We base our methodology on a Lasso-pe
nalized M-estimator which\, despite non-convexity\
, may be solved by a polynomial-time algorithm suc
h as coordinate or gradient descent. Our results p
rovide theoretical guarantees on asymptotic normal
ity of the new estimators and may be used for vali
d hypothesis testing and variable selection. These
results are achieved under a sparsity condition r
elating the number of non-zero loadings\, sample s
ize\, dimensionality of the covariance matrix and
spectrum of the covariance matrix. This talk is ba
sed on joint work with Sara van de Geer.
LOCATION:Seminar Room 1\, Newton Institute
CONTACT:INI IT
END:VEVENT
END:VCALENDAR