Towards an advanced system for real-time event detection in high-volume data streams

This paper presents an advanced system for real-time event detection in high-volume data streams. Our main goal is to provide a system, which can handle high-volume data streams and is able to detect events in real-time. Additionally, we perform further steps, such as classifying and ranking events with retrospective analysis. To solve this task we take advantage of a high-performance database system for semi-structured data and extend it with the functionality of continuous querying. The combination of executing queries on the incoming data stream and fast queries on the historical datasets is used as a powerful tool for developing an event detection and information system. Furthermore, we define several event features for improving event classification and for discovering parallelisms, relations, duration, and coherences of events.