201 search results - page 3 / 41 » Exploiting Unlabeled Data for Improving Accuracy of Predicti... |
PAKDD
2005
ACM
13 years 11 months ago
2005
ACM
The usefulness of the results produced by data mining methods can be critically impaired by several factors such as (1) low quality of data, including errors due to contamination, ...
NIPS
2008
13 years 7 months ago
2008
The cluster assumption is exploited by most semi-supervised learning (SSL) methods. However, if the unlabeled data is merely weakly related to the target classes, it becomes quest...
KDD
2009
ACM
2009
ACM
Quantification and semi-supervised classification methods for handling changes in class distribution
14 years 6 months ago
In realistic settings the prevalence of a class may change after a classifier is induced and this will degrade the performance of the classifier. Further complicating this scenari...
KDD
2006
ACM
14 years 6 months ago
2006
ACM
Long-term search history contains rich information about a user's search preferences. In this paper, we study statistical language modeling based methods to mine contextual i...
BMCBI
2010
13 years 5 months ago
2010
Background: Recently, supervised learning methods have been exploited to reconstruct gene regulatory networks from gene expression data. The reconstruction of a network is modeled...