If your data can be "cut" and unrolled at a specific boundry then you
can rotate/translate the data away from the boundry. For example if
your data crosses the 0 degree boundry but not the -180/+180 boundry
then you don't need to do anything, if it crosses the -180/+180
boundry but not the 0 degree boundry then you can translate the data
for that cluster/variable with the following:
if old >= 0 then new := old - 180
else if old < 0 then new := old + 180.
and then translate the results back when you are finished. This
allows you to compute the means of the data correctly. You only need
to do this for data clusters that cross the 180 degree boundry.
If however your data is actually distributed entirely on a circular
range, I don't think this technique will work as it will artifically
increase the distance between elements that were previously close.
-Jason
"Carl Frelicot" <[EMAIL PROTECTED]> writes:
>I face the problem of clustering one-dimensional data that can range in a
>circular way. Does anybody knows the best way to solve this problem with no
>aid of an additional variable ? Using a well-suitable trigonometric
>transform ? Using an ad-hoc metric ?
>Thanks.
--
-------------------------------------------------------------------------------
J. [EMAIL PROTECTED] http://www.cs.ubc.ca/~harrison
Graduate Motto: Free-time with guilt. ftp://ftp.cs.ubc.ca/pub/local/quotes
===========================================================================
This list is open to everyone. Occasionally, less thoughtful
people send inappropriate messages. Please DO NOT COMPLAIN TO
THE POSTMASTER about these messages because the postmaster has no
way of controlling them, and excessive complaints will result in
termination of the list.
For information about this list, including information about the
problem of inappropriate messages and information about how to
unsubscribe, please see the web page at
http://jse.stat.ncsu.edu/
===========================================================================
