Support vector machines for spam categorization
IEEE Transactions on Neural Networks
Hi-index | 0.00 |
We studied the objective of spam based on a financial profit truth – the cost for sending spam against anti-spam techniques should be little than the return from the negligible response from the recipients. Spammers have to leave some real contact information in the spam for the recipients to touch them easily, no matter what methods they use to fight against anti-spam techniques. In this paper, we present a method to automatically identify such contact information entities in spam, and build an online blocklist for the spam filters to classify spam, especial unsolicited commercial email (UCE).