Re: ? most representative vector of a cone

On Tue, 19 Sep 2006 03:11:56 GMT, "Cheng Cosine"
<acosine@xxxxxxxxxxxx> wrote:

Hi:

Given a set of vectors that are nearly parallel to each other but with
different lengths.

To measure how close these vectors are to be parallel to each other, say
they are confined

in a cone with small angle theta.

Now can we analytically choose a vector that is most representative to
this set of vectors

so that the max difference dot prodcut of an arbitrary vector to any of this
set of vectors

with the dot product of this representative vector with this arbitrary
vector is minimized?

Thanks,
by Cheng Cosine
Sep/19/2k6 NC


Not sure exactly how you are proposing to do that, but remember that
you need to make them unit vectors if you wish to compare directions
without being influenced by length.

--Lynn
.

Quantcast