[rrd-users] Aberrant Behavior Detection tuning

Praveen P pra_ve_en at yahoo.com
Wed Aug 9 05:10:26 MEST 2006


Can someone please explain how to tune the seasonal
coefficient adaptation and seasonal deviation
adaptation parameters for aberrant behavior detection

By default, gamma is set to the same value as alpha. I
understand that the Alpha value is derived based on
the "% weight given to the last n number of samples".
So does this mean that gamma is also a weight for a
set of seasonal co-efficients or does it have a one to
one correspondence between the current prediction and
a single value one seasonal period in the past ?

I was trying to see if the algorithm can 'learn' about
the deviations in the pattern that occurs at a fixed
point in every seasonal cycle. For example, if we were
monitoring web tv traffic, we would like the algorithm
to learn the fact that every monday there will be a
sudden increase in the traffic at the time of the
football telecast (assuming the seasonal period to be
one week). Is this possible currently ?

Thanks, Praveen.

Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 

Unsubscribe mailto:rrd-users-request at list.ee.ethz.ch?subject=unsubscribe
Help        mailto:rrd-users-request at list.ee.ethz.ch?subject=help
Archive     http://lists.ee.ethz.ch/rrd-users
WebAdmin    http://lists.ee.ethz.ch/lsg2.cgi

More information about the rrd-users mailing list