On 5/8/06, Erik Volz <[log in to unmask]> wrote:
<-- good information snipped, but then... -->
> Most of the other
> papers concern statistical models of networks rather than descriptive
> metrics. I suspect they would take you rather far afield.
Be careful. Any "descriptive metric" assumes some range of appropriate
statistical model for which it conveys knowledge, and not just information,
often being able to be derived as an estimator from such a model (sometimes
that is the only derivation.
(classic simple example: using a mean or median (or any measure of
"centrality") for a bi-modal ("2-hump") distribution -- you get information,
but usually the knowledge you want, which is the "most likely" or "common
tendency", goes missing).
best,
-tony
[log in to unmask]
Muttenz, Switzerland.
"Commit early,commit often, and commit in a repository from which we can
easily
roll-back your mistakes" (AJR, 4Jan05).
--
best,
-tony
[log in to unmask]
Muttenz, Switzerland.
"Commit early,commit often, and commit in a repository from which we can
easily
roll-back your mistakes" (AJR, 4Jan05).
|