Publication at Faculty of Mathematics and Physics |
2009
Abstract
This paper describes POS tagging experiments with semi-supervised training as an extension to the (supervised) averaged perceptron algorithm, first introduced for this task by Collins02. Experiments with an iterative training on standard-sized supervised (manually annotated) dataset (10^6 tokens) combined with a relatively modest (in the order of 10^8 tokens) unsupervised (plain) data in a bagging-like fashion showed significant improvement of the POS classification task on typologically different languages, yielding better than state-of-the-art results for English and Czech (4.12 % and 4.86 % relative error reduction, respectively; absolute accuracies being 97.44 % and 95.89 %).