> -----Original Message-----
> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On
> Behalf Of Reggie Bautista
...
> On that social network analysis... I remember someone talking
> about judging
> the strength of argument in an email based on the number of replies. Was
> that you? If so, how do filter for people like me who go through phases
> where I'm too busy to look in on the list for a few days/weeks/months,
> followed by phases where I read and post every day? I don't have any
> suggestions, mind you, I'm just curious if you've even considered that
> angle.
Most of the analysis I'm doing is short-term. The longest time period for
almost anything I do is 30 days, so it doesn't really care what you do over
the long run.
However, it's not so much the number of replies as how the community behaves
following your posts. One of the most significant factors, it seems, is
what happens to the number of people posting messages (in any thread) after
your posts. On the other hand, threading is a bit of a processing
challenge, so I'm not yet doing a thorough job of measuring thread-related
evidence.
The graphs I posted actually weren't a big part of what I'm working on.
They actually did cover longer periods of time and tried to measure the
links between people. In that context, coming and going as you describe
would have an impact. That's why none of these approaches stands alone. In
fact, I'm working on getting many factors into the system, treating it as a
Boolean state space and using an evolutionary computing approach to try to
find the most sigificant factors and any attractor states that might exist.
Nick
