In the paper, we give an overview of several approaches that we use to analyze “spam” (undesired bulk advertising) and the credentials of senders of spam, for the purpose of automatic detection. We use some of these approaches for discrimination between stolen account credentials and “spam-bots” (accounts opened purely for the purpose of distributing spam), since the methods for handling spam senders differs by type of account. Other approaches described in this paper we use to automatically detect classes of spam messages, assisting spam analysts in their work to find the messages, so that the messages may be automatically deleted.

batura.pdf613.26 KB