Algorithm Based On Bayesian Filtering, Anti-spam Technology And Its Implementation

Junk Email becomes a serious problem recently. In order to resolve the issueultimately, it needs to make great efforts from three aspects: legislation,organization and anti-spam technology. The author works from the way ofanti-spam technique and based on researching a great deal of related anti-spamskills, choosing one method named "Bayesian Filter arithmetic" as the centraltask. After collecting a lot of email samples, the author investigates the arithmeticdeeply and implements it with VC++. The product named "AntiSpam", as adesktop application, it works as a proxy between the email client and the emailserver to filter the spam.The application fully meets with user's personal requirements with betterself -adaptability. It analysis the user's current email samples and recognize theattributes of spam. In this way, it will filter the new messages when it arrives.Both English and Chinese are supported. As a solution for anti-spam on desktop,it can lighten the effect of junk email.The AntiSpam composes of six parts: email proxy, email parser, email filter,Chinese word split, parameters configuration and log analysis. This paperimplements the arithmetic for Bayesian Filter with C++ and choose differentparameters to improve the performance. The email filtering system mainly usedthis junk email recognizing mechanism to work. Another part, the email proxymodule using "Winsock", works like a proxy, it will filter the junk email afterreceiving a new message from email server. It integrates with the user's currentemail client software and can be easily deployed and without much moreadjustment. Both of the two parts are the key of this paper.