Implementation of mail filtering based on multi-attribute group decision making

The judgment of spam usually has some uncertainties, so it is easy to be misjudged, which lead to great loss. Aiming at the situation of misjudgment to mail filter presently, a method of spam filtering based on multi-attribute group decision making is presented in this paper. This kind of mail filtering is considered as processing of uncertain information. A reasonable comprehensive evaluation value is computed as a membership degree by integrating ordered weighted averaging (OWA) operator and combination weighted averaging (CWAA) operator. Fuzzy support vector machines (FSVM) classification algorithm is used to classify mails. The experiments show that this method of mail filtering improves the Recall and Precision of spam filtering to some extent.