Junk E-mail Filtering Method Based on Data Mining

The authors analyze the essence of some existing algorithms of junk E-mail filtering rules and give the way to represent an E-mail as a transaction. After the datasets are expressed as the transaction sets, the frequent feature sets can be mined from the datasets by means of FP-Tree algorithm. Finally, the problem about how to express the frequent feature sets as rules and how to apply the rules is also discussed.