Data Intensive Computing
CLIQUE — Real-time Event Detection and Visualization
Principal Investigators: Bill Pike, Shawn Bohn
Challenge
Provide analysts with real-time streaming data about events of interest with sufficient context to make a response decision.
CLIQUE
Approach
- Extend event processing techniques to support detection and contextualization of complex event patterns in streaming data.
- Implement new event-caching mechanism for tracking evolving patterns.
- Provide visual monitoring interfaces that support human identification of event status.
Impact
Reduced burden on human analysts to discover complex event sequences, such as sophisticated cyber attacks.
Collaborations
- Stanford University
- U.S. Computer Emergency Readiness Team
Accomplishments
- Prototype application for detecting events of interest in streaming cyber data. (May 08)
- Connectivity to Stanford University's large-volume data visualization system. (Jun 08)
- Implementing Very Fast Machine Learning (VFML) to perform pattern identifications. VFML provides classification and clustering in constant time, regardless of data set size.