367 search results - page 2 / 74 » A weighted subspace approach for improving bagging performan... |
IDA
2007
Springer
13 years 11 months ago
2007
Springer
Random forests are one of the best performing methods for constructing ensembles. They derive their strength from two aspects: using random subsamples of the training data (as in b...
PAKDD
2009
ACM
14 years 4 days ago
2009
ACM
It is a challenging and important task to retrieve images from a large and highly varied image data set based on their visual contents. Problems like how to fill the semantic gap b...
IJCV
2006
13 years 5 months ago
2006
Subspacefacerecognitionoftensuffersfromtwoproblems:(1)thetrainingsamplesetissmallcompared with the high dimensional feature vector; (2) the performance is sensitive to the subspace...
SSPR
2000
Springer
13 years 9 months ago
2000
Springer
To improve weak classifiers bagging and boosting could be used. These techniques are based on combining classifiers. Usually, a simple majority vote or a weighted majority vote are...
TSP
2008
13 years 5 months ago
2008
A new blind subspace-based channel estimation technique is proposed for direct-sequence code-division multiple access (DS-CDMA) systems operating in the presence of unknown wide-se...