MAIDS: Mining Alarming Incidents from Data Streams

Real-time surveillance systems, network and telecommunication systems, and other dynamic processes often generate tremendous (potentially infinite) volume of stream data. Effective analysis of such stream data poses great challenges to database and data mining researchers, due to its unique features, such as single-scan algorithm, multi-dimensional online analysis, fast response time, etc. In this paper we propose to demonstrate our stream data mining system, MAIDS--Mining Alarming Incidents from Data Streams, which is a project supported by U.S. Office of Naval Research and National Science Foundation, jointly developed by Automated Learning Group, NCSA and Department of Computer Science, the University of Illinois at Urbana-Champaign. By integration of our most recent research results on stream data analysis, we have successfully developed the MAIDS system within the D2K data mining framework [19] with the following distinct features: (1) a tilted time window framework and multi-res...

• Type: Conference
• Year: 2004
• Where: SIGMOD
• Authors: Y. Dora Cai, David Clutter, Greg Pape, Jiawei Han, Michael Welge, Loretta Auvil
• Keywords: Database, Sensor Stream Data, SIGMOD 2004, Stream Data Analysis, Stream Data Mining