Generalized tri-training

Generalized tri-training is a semi-supervised learning algorithm, exploiting both labeled and unlabeled data. The algorithm was implemented in Orange 2.0b for Mac OS X. The code is available here.


Soegaard, Anders; Rishoej, Christian. 2010. Semi-supervised dependency parsing using generalized tri-training. The 23rd International Conference on Computational Linguistics (COLING). Beijing, China.

Blå linie
Njalsgade 140-142, bygn. 25, DK-2300 KBH S
Tlf: +45 35329090 - Fax: +45 35329089
Valid HTML 4.01 Strict