Robust anomaly detection and regularized domain adaptation of classifiers with application to internet packet-flows

摘要: Sound, robust methods identify the most suitable, parsimonious set of tests to use with respect prioritized, sequential anomaly detection in a collected batch sample data. While focus is on detecting anomalies network traffic flows and classifying into application types, are also applicable other classification settings, including email spam, (e.g. credit card) fraud detection, imposters, unusual event (for example, images video), host-based computer intrusion equipment or complex system failures, as well anomalous measurements scientific experiments.