Sarang,
I am Sarang Shrivastava, an open source enthusiast. This year in the
Gsoc
2015 a series of events happened after which I landed up here with SA.
I made a proposal for SA "Building a statistical classifier plugin for
SA"
Thanks for applying to GSoC, and thanks to Kevin for accepting
mentorship.
1) Can you guys just tell me the reason why ANN's are not included as a
filter with SA?
Nobody implemented it so far.
(although GA algorithms are used offline in the SpamAssassin project
to assign score points to rules)
2)If anyone amongst you tried to implement it, what were the
difficulties
you faced ?
Don't know. It might be worth taking a look at the CRM114 classifier,
which implements a number of methods. The CRM114 can be used as a
plugin to SpamAssassin, or can be called from Amavis and results
combined with those from SpamAssassin. We had good results from CRM114,
and some friends of mine are very enthusiastic about it. Certainly
it is a good complement to SpamAssassin's naive bayes classifier,
3)Can anyone of you guide me somewhere that makes me beleive more that
yes
I am moving in a right direction ?
I think it is definitely worth trying new approaches and compare
results.
Even if it may turn out to be less successful, learning and
experimenting
is valuable.
Mark
