Comment Re:Mailing lists / newsletters (Score 3, Insightful) 326
According to the paper this article is based on, the algorithm is effective against messages with multiple recipients in the To: or Cc: headers. This means that messages coming from slashdot-headlines@newsletters.osdn.com would probably be in the unclassifiable half. Indeed, a good chunk of spam these days would be unclassifiable according to this algorithm.
However, the whitelist that this algorithm generates would still be valid. To me, this is the real strength of the algorithm, to be able to generate a white list with no input on my part.
However, the whitelist that this algorithm generates would still be valid. To me, this is the real strength of the algorithm, to be able to generate a white list with no input on my part.
