System and method for detecting spam using clustering and rating of e-mails

Abstract

Disclosed are systems, methods and computer program products for detection of spam. In one example, a system receives electronic messages and attempts to classify the messages as legitimate or spam messages. For an unknown message, the system obtains its metadata including hash sum of the message and sender's IP address. The system then places the metadata of the unknown messages into one cluster of a plurality of clusters based on degree of similarity between hash sums of different unknown messages. The system then rates each unknown message in accordance with a rating of the cluster, wherein the rating of the cluster is based, at least, on a number of similar hash sums of unknown messages received from different addresses of message senders contained in said cluster. Finally, the system classifies unknown messages as legitimate or spam based on the message rating.