Detecção de Spammers na Rede de Origem

Pedro Henrique B. Las CasasDorgival GuedesJussara M. AlmeidaArtur ZivianiHumberto T. Marques-Neto

The volume of unsolicited messages (spam) sent over the Internet representsmore than 85% of all e-mails. Even with the evolution of the filteringtechniques such as the analysis of the message content and the blocking of IPs,network resources are wasted given that such a filtering is usually performed atthe e-mail destination server. This paper proposes a method for detecting spammersin the origin network using some metrics which do not require inspectionof message contents. This method uses a supervised classification technique thathas been applied to the two real-world datasets from a Brazilian broadband Internetservice provider. Results show that the adopted method is efficient, beingable to correctly identify most spammers still in the origin network. In this way,network resources are saved because of the reduced number of spams in transitthat likely would be discarded at their destination.

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato:
     Mantida por: