E-mail Filtering System Research And Design

As a kind of important service on the Internet—Email, provides a kind of important communication means for people. Because of the defect on the E-mail principle, It have caused more and more Unsolicited Bulk Email that have already caused people's great attention. The mail filtering technology has already become one of the focuses of technology researching at present.This paper has designed a kind of mail filtering system based on Linux platform, removing the virus in mails through the virus-scanning engine, adopting the categorized algorithm of text based on Vector Space Model to classify the mail according to the content of the mail. Thus it prevents the Spam from causing the harmful effects to the mail server and mail user.The paper has studied the principle and related protocols of the E-mail at first, introduced the present condition and the harm of the spam and various kinds of anti-spam techniques and the related products. After that the paper has analyzed the characteristic of each kind of text classification algorithms, has carried on the analysis to the reason why the sort precision of Vector Space Method classification is not high, improving the algorithm from two aspects: the characteristic extracting and the power computing, increasing the Vector Space Method's classification precision effectively, and has used this improved algorithm as the classified algorithm of the mail filtering system which this paper designed. The paper has analyzed the questions of mail decoding and mail text information standardization which the mail filtering involves, has realized the Chinese text participle algorithm based on stopped words table, has carried on research on characteristic extracting and the power computing and the classified valve's value computing which are the correlation questions of Space vector method.The mail filtering system which the paper designed using multistage filtering model, has realized multistage filters to classify Emails based on the rule and the text content, can discriminate the junk mail from normalmail effectively, has higher application value.