The Research & Design Of Spam Mail Determinator

The junk mail enormously increase its proportion from 7% at 2001 to be anxious 51% current .now, which become one of the biggest public damage on global internet society.In allusion to the overrun junk mails on Internet, the first part of this paper collect the junk mail filtering technique documentations and information, and develop the deep analysis. Combination of my current job, I also conclude the existing problems in current anti-junk mail systems.This thesis proposes a set of speed,effective and intelligent junk mail determinators, which not only adapt the main stream junk mail filtering techniques and method, such as: Bayesian, RBL filtering; but also implement the combination of other security system intelligent techniques into our platform, for example, fingerprint analysis techniques in anti virus system,semantic analysis and neural network techniques in content filtering system. At the same time, in this thesis, based on hierarchical filter system structure, data hierarchy flow design, the problems in anti junk mail system, such as slow mail filtering, are solved effectively.However, there still exist in our thesis some difficult problems remained incompletely solved, .e.g. how to filter graphical junk mail effectively? There still need further research in study and work.