One important problem that costs big companies billions
of dollars annually because of employees loss of productivity is
represented by the high amount of spam entering their accounts. An
evolutionary model for filtering spam is proposed. It constructs a
set of rules from a training set and them uses them to

Spam has very rapidly grown to be more than just an unpleasant
presence. It currently costs big companies billions of dollars
annually because of employees loss of productivity. An
evolutionary model for spam detection is proposed. It evolves a
set of rules that induced the membership of e-mails from a
training set either to the 'spam' or the 'ham' category. Rules are
self adapting for their corresponding category through the means
of a special evolutionary metaheuristics called Genetic
Chromodynamics. Detected rules are used to classify previously
unseen e-mails. Proposed technique may be used for more general
purposes as system identification, text categorization etc.