1175 search results - page 4 / 235 » Experimental perspectives on learning from imbalanced data |
ICML
2003
IEEE
14 years 7 months ago
2003
IEEE
When the training instances of the target class are heavily outnumbered by non-target training instances, SVMs can be ineffective in determining the class boundary. To remedy this...
ICPR
2008
IEEE
14 years 21 days ago
2008
IEEE
This paper presents a new learning approach for pattern classification applications involving imbalanced data sets. In this approach, a clustering technique is employed to resamp...
NCA
2007
IEEE
13 years 5 months ago
2007
IEEE
Learning from imbalanced data occurs frequently in many machine learning applications. One positive example to thousands of negative instances is common in scientific applications...
SIGKDD
2008
13 years 6 months ago
2008
This paper presents our solution for KDD Cup 2008 competition that aims at optimizing the area under ROC for breast cancer detection. We exploited weighted-based classification me...
ICPR
2006
IEEE
14 years 7 months ago
2006
IEEE
A two-class imbalanced data problem (IDP) emerges when the data from majority class are compactly clustered and the data from minority class are scattered. Though a discriminative...