1126 search results - page 86 / 226 » Constraint Classification: A New Approach to Multiclass Clas... |
KDD
2006
ACM
14 years 8 months ago
2006
ACM
Kernel machines have been shown as the state-of-the-art learning techniques for classification. In this paper, we propose a novel general framework of learning the Unified Kernel ...
ICCV
2009
IEEE
13 years 5 months ago
2009
IEEE
Hybrid generative-discriminative techniques and, in particular, generative score-space classification methods have proven to be valuable approaches in tackling difficult object or...
CVPR
2010
IEEE
13 years 11 months ago
2010
IEEE
A clustering framework within the sparse modeling and dictionary learning setting is introduced in this work. Instead of searching for the set of centroid that best fit the data, ...
CVPR
2006
IEEE
14 years 10 months ago
2006
IEEE
We consider the general problem of learning from both labeled and unlabeled data. Given a set of data points, only a few of them are labeled, and the remaining points are unlabele...
BMCBI
2005
13 years 7 months ago
2005
Background: Structure-dependent substitution matrices increase the accuracy of sequence alignments when the 3D structure of one sequence is known, and are successful e.g. in fold ...