Making large-scale support vector machine learning practical
Advances in kernel methods
Combining Labeled and Unlabeled Data for Text Classification with a Large Number of Categories
ICDM '01 Proceedings of the 2001 IEEE International Conference on Data Mining
Reducing multiclass to binary: a unifying approach for margin classifiers
The Journal of Machine Learning Research
ICDAR '05 Proceedings of the Eighth International Conference on Document Analysis and Recognition
Solving multiclass learning problems via error-correcting output codes
Journal of Artificial Intelligence Research
Dynamic multiple fault diagnosis: mathematical formulations and solution techniques
IEEE Transactions on Systems, Man, and Cybernetics, Part A: Systems and Humans - Special section: Best papers from the 2007 biometrics: Theory, applications, and systems (BTAS 07) conference
Exploiting System Knowledge to Improve ECOC Reject Rules
ICPR '10 Proceedings of the 2010 20th International Conference on Pattern Recognition
| Hi-index | 0.00 |
The original task of a multiclass classification problem can be decomposed using Error Correcting Output Coding in several two-class problems which can be solved with dichotomizers. A reject rule can be set on the classification system to improve the reliability of decision through an external threshold on the decoding outcomes before the decision is taken. If a loss-based decoding rule is used, more can be done to make such external scheme works better introducing a further reject stage in the system. This internal approach is meant to single out unreliable decisions for each classifier in order to proficiently exploit the properties of loss decoding techniques for ECOC as proved by experimental results on popular benchmarks.