Método Heurístico para Rotular Grupos em Sistema de Detecção de Intrusão baseado em Anomalia

Hermano PereiraEdgard Jamhour

The intrusion detection systems are part of security suite necessary for environment protection that contains information available on the Internet. Among these systems it is highlighted the unsupervised learning, as they are able to extract environment models without prior knowledge concerning the occurrence of attacks among the collected information. A technique used to create these models is the data clustering, where the resulting clusters are labeled either as normal or as attack (in anomalous case). This paper proposes a heuristic method for labeling clusters, where the false positive rates achieved during experiments were significantly lower compared to the methods described in related work.

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: