1126 search results - page 31 / 226 » Constraint Classification: A New Approach to Multiclass Clas... |
ICCV
2009
IEEE
13 years 5 months ago
2009
IEEE
Multi-label learning is useful in visual object recognition when several objects are present in an image. Conventional approaches implement multi-label learning as a set of binary...
JCDL
2005
ACM
14 years 1 months ago
2005
ACM
Automatic text classification is an important operational problem in digital library practice. Most text classification efforts so far concentrated on developing centralized solut...
KDD
2008
ACM
14 years 8 months ago
2008
ACM
We propose a family of novel cost-sensitive boosting methods for multi-class classification by applying the theory of gradient boosting to p-norm based cost functionals. We establ...
KDD
2002
ACM
14 years 8 months ago
2002
ACM
Support vector machines (SVMs) excel at two-class discriminative learning problems. They often outperform generative classifiers, especially those that use inaccurate generative m...
ICIAP
2009
ACM
14 years 8 months ago
2009
ACM
Abstract. Score functions induced by generative models extract fixeddimensions feature vectors from different-length data observations by subsuming the process of data generation, ...