The ability to mine data represented as a graph has become important in several domains for detecting various structural patterns. One important area of data mining is anomaly det...
The ability to mine data represented as a graph has become important in several domains for detecting various structural patterns. One important area of data mining is anomaly det...
William Eberle, Lawrence B. Holder, Jeffrey Graves
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...
—In this paper we present graph-based approaches to mining for anomalies in domains where the anomalies consist of unexpected entity/relationship alterations that closely resembl...
We consider the task of performing anomaly detection in highly noisy multivariate data. In many applications involving real-valued time-series data, such as physical sensor data a...