Spam-mail Filtration Studies And Realizes

Spam, acting as the carrier of business advertisements, the virus, the malicious programs or some sensitive mails, are more and more fiercely threatening the safety of the computer systems and the common lives of people, making service provider and client have to spend a large amount of time and energy in processing these spam, and wasting a great deal of time and money. Therefore, Anti-spam has become an international, significant and practical topic by now.This paper suggests a sort of spam filter system framework base on mail contents via researching the spam filter technologies and arithmetic in existence and making a summary of the strongpoint and shortage of them. This method starts with the content characters of the mail, studies on spam mails by Bayes study means, and gets original spam mails character concourse as the base of the filter. Because mail contents is different with time going on and change of subject and it can't update keyword in time by setting prior keyword, filter efficiency is not ideal. Our system adopts the method of keyword self-update, which improves filter's adaptability effectively. With the increase of the kind of mails, the expansibility of the system filter becomes more and more important. When needed, new function can be added to the filter to filter other type of mails , such as image mail.this paper implements a prototype of anti-spam filter system.The whole filter process is transparent to the users, having no affect on users'normal network action.the system applies deleting-repetition to the mail message before making judgment, in order to wipe off an amount of spam and reduce fluxes.In the process of judgment, the system uses Bayes judgment means, reducing miscarriage of justice on spam mails and normal mails.The system can communicate the study result in this network segment to the system in another network segment by communication module. And the study result is picked up by integrative study module in the system as correlative rules.In addition, the system supports user-friendly configuration and management user interface, and allows the user to add in filtration characteristic, making the system more flexible. Some important system operations can be written in the log, making the user query later and monitor the system easily.In essence, the relation of the creation and filter of spam is just like spear and shield, both making progress ceaselessly. There isn't and won't a good filter system in function for good. It is especially important to make research on the study ability of the filter system, and it is what the project will do in the future.