I see you're a gmail user. You have no idea how much spam Gmail is
rejecting or discarding before it ever gets anywhere near your inbox.
It doesn't matter how much spam is rejected overall: what matters is
how much ham is rejected, and how early the spam can be rejected, and,
in the case of the phone, how much more tightly we can draw the
filtering compared to a desktop machine. ...
Really, you cannot assume that the mail you get at your gmail account is
typical of anything. You also have no idea how much spam Gmail doesn't
even put in your spam folder, based on content analysis they didn't tell
you about.
This can be easily quantified in the case of naive Bayesian
classifiers, by looking at the entropy gain of each signal, and doing
the usual sort of threshold picking analysis.
Um, have you ever talked to people who run large mail systems about the
way their spam filtering really works? Many of us here have done so,
and it's a lot more complicated than it might seem.
R's,
John
_______________________________________________
Endymail mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/endymail
