[ Date Index ] [ Thread Index ] [ <= Previous by date / thread ] [ Next by date / thread => ]
On Sun, Feb 8, 2009 at 9:01 PM, Simon Waters wrote: > Wondered if anyone has seen, or used Bayesian filtering for other > classifications of documents other than spam? > > I guess "not wanted" is often equivalent to "spam", but it occurred to > me that you could train it for any criteria you want, such as > "interesting" or "well written" etc. Not entirely sure if it was Bayesian, but I think they've used it to classify documents written by someone, e.g. to see whether a recently discovered 17-th century play could have been written by Shakespeare. In all these cases, Markovian discrimination might give better results though. Martijn. -- The Mailing List for the Devon & Cornwall LUG http://mailman.dclug.org.uk/listinfo/list FAQ: http://www.dcglug.org.uk/linux_adm/list-faq.html