Prevention Mechanisms, Based On The Full-text Filtering Spam

With the fast development of Internet, the use of the E-mail becomes one of the most fast and economic method of communication in people's daily life. But the spam question increased day by day which forces users spend much time and energy to deal with the spam. So how to get the demanded mail exactly and decrease the disturber of the spam has become a social problem.The methods which based on separate word searches of content and regular expressioncan filter spam which accord with the regular. But at the same time, disadvantage is also obviously, accuracy is not good enough and can not deal with some multi-vocals, interpretations-words. The problem of data redundancies and veracity is seriously.The thesis bring forward one arithmetic based on split word, token the label of word's attribute and max matching word positive to recede the shortcomings. Next, the thesis discuss the develop way of the future spam technology, just like fuse clustering and vsm in filter which has better effect.The thesis mainly discussed how to keep away use full filter which depend on what reason bring the spaim regular expression on spam filter, full searches and semantic analyze, By using mature technology , succeed in design and realize one spam filter system in Wan which has gateway transaction and spam dispose. The system using spite words and filter technology, clustering technology succeed divided the E-mail into two category: mail legality and spam.