The two main challenges typically associated with mining data streams are concept drift and data contamination. To address these challenges, we seek learning techniques and models ...
One of the most challenging problems in data manipulation in the future is to be able to e ciently handle very large databases but also multiple induced properties or generalizatio...
Combining retrieval results from multiple modalities plays a crucial role for video retrieval systems, especially for automatic video retrieval systems without any user feedback a...
We have been developing a data mining (i.e., knowledge discovery) framework, MADAM ID, for Mining Audit Data for Automated Models for Intrusion Detection [LSM98, LSM99b, LSM99a]. ...
We explore in this paper an effective sliding-window filtering (abbreviatedly as SWF) algorithm for incremental mining of association rules. In essence, by partitioning a transact...