We introduce a novel active-learning scenario in which a user wants to work with a learning algorithm to identify useful anomalies. These are distinguished from the traditional st...
In this paper, we study the problem of anomaly detection in high-dimensional network streams. We have developed a new technique, called Stream Projected Ouliter deTector (SPOT), t...
One of the primary issues with traditional anomaly detection approaches is their inability to handle complex, structural data. One approach to this issue involves the detection of...
If efficient network analysis tools were available, it could become possible to detect the attacks, anomalies and to appropriately take action to contain the attacks. In this paper...
Seong Soo Kim, A. L. Narasimha Reddy, Marina Vannu...
—Traffic anomaly detection has received a lot of attention over recent years, but understanding the nature of these anomalies and identifying the flows involved is still a manu...