Cis6930-clus_epid - Class
presenta+on
on
clustering


Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Class
presenta+on
on
clustering
 and
Epidemic
modeling
 ‐Udayan
Kumar
 Clustering
 •  Allows
unsupervised
classifica+on
of
input
 data
into
different
classes.
 •  Useful
for
grouping
users
based
on
their
 usage.
For
example
separa+ng
Regular
users
 from
visitors
 •  K‐means,
PAM,
Hierarchical

clustering
and
 several
more.
 •  R
and
Matlab
have
god
collec+on
of
libraries
 for
clustering.
 Clustering
 •  I
have
used
it
to
cluster
regular
users

and
 visitors
at
a
loca+on
 •  Clustering
done
on
+me
spend,
number
of
 sessions
and
dis+nct
days
of
login

 Epidemic
Modeling
 •  How
to
study
connec+vity
in
a
Dynamic
 encounter
graphs?
 •  The
challenges
are
:

 1.  Network
connec+vity
changes
with
+me
 2.  Measuring
characteris+cs
of
the
network
 3.  How
to
compare
the
performance
of
your
 protocol
to
the
best/base
case
 Epidemic
Modeling
 •  Epidemic
rou+ng
is
a
controlled
flooding.
 •  Message
is
given
to
every
encountered
node
if
 that
node
has
not
already
received
the
 message.
 •  Every
receiver
becomes
the
message
sender
 too.

 •  This
allows
us
to

measure
reachability,
delay
 and
overhead.
 ...
View Full Document

This note was uploaded on 05/27/2011 for the course CIS 4930 taught by Professor Staff during the Spring '08 term at University of Florida.

Ask a homework question - tutors are online