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:Estimating network edge probabilities by neighborh
ood smoothing - Elizaveta Levina (University of Mi
chigan)
DTSTART;TZID=Europe/London:20160714T113000
DTEND;TZID=Europe/London:20160714T120000
UID:TALK66753AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/66753
DESCRIPTION:Co-authors: Yuan Zhang (Ohio State Universit
y)\, Ji Zhu (University of Michigan)

The problem of estimating probabilities of net
work edges from the observed adjacency matrix has
important applications to predicting missing link
s and network denoising. It has usually been addr
essed by estimating the graphon\, a function that
determines the matrix of edge probabilities\, but
is ill-defined without strong assumptions on the
network structure. Here we propose a novel compu
tationally efficient method based on neighborhood
smoothing to estimate the expectation of the adja
cency matrix directly\, without making the strong
structural assumptions graphon estimation require
s. The neighborhood smoothing method requires lit
tle tuning\, has a competitive mean-squared error
rate\, and outperforms many benchmark methods on
the task of link prediction in both simulated and
real networks.

Related Links
LOCATION:Seminar Room 1\, Newton Institute
CONTACT:INI IT
END:VEVENT
END:VCALENDAR