70 search results - page 4 / 14 » Boosting Based on a Smooth Margin |
IDA
2002
Springer
14 years 9 months ago
2002
Springer
This paper introduces a strategy for training ensemble classifiers by analysing boosting within margin theory. We present a bound on the generalisation error of ensembled classifi...
108
Voted
IDEAL
2010
Springer
14 years 6 months ago
2010
Springer
Based on studies and experiments on the loss term of SVMs, we argue that 1-norm measurement is better than 2-norm measurement for outlier resistance. Thus, we modify the previous 2...
75
Voted
ICPR
2008
IEEE
15 years 10 months ago
2008
IEEE
AdaBoost and support vector machines (SVM) algorithms are commonly used in the field of object recognition. As classifiers, their classification performance is sensitive to affect...
ICMLA
2004
14 years 11 months ago
2004
AdaBoost rarely suffers from overfitting problems in low noise data cases. However, recent studies with highly noisy patterns clearly showed that overfitting can occur. A natural s...
99
Voted
ML
2002
ACM
14 years 9 months ago
2002
ACM
We examine linear program (LP) approaches to boosting and demonstrate their efficient solution using LPBoost, a column generation based simplex method. We formulate the problem as...