406 search results - page 17 / 82 » New ensemble methods for evolving data streams |
ISBI
2011
IEEE
13 years 13 days ago
2011
IEEE
We develop a new algorithm to segment the hippocampus from MR images. Our method uses a new classifier ensemble algorithm to correct segmentation errors produced by a multi-atlas...
IJCAI
2003
13 years 10 months ago
2003
Ensemble methods like bagging and boosting that combine the decisions of multiple hypotheses are some of the strongest existing machine learning methods. The diversity of the memb...
ASC
2008
13 years 8 months ago
2008
: Discovering interesting patterns or substructures in data streams is an important challenge in data mining. Clustering algorithms are very often applied to identify single substr...
ICDE
2007
IEEE
14 years 3 months ago
2007
IEEE
The ability to query data streams is of increasing importance and has been identified as a crucial element for modern organizations and agencies. This paper identifies a class o...
EDBT
2008
ACM
14 years 8 months ago
2008
ACM
There has been much recent interest in on-line data mining. Existing mining algorithms designed for stored data are either not applicable or not effective on data streams, where r...