Sepe: A POS Tagger for Spanish

摘要

We describe a part-of-speech tagging system specially designed to tag Spanish texts using small linguistic resources. Nevertheless, the tagger obtains encouraging results. We have found and exploited useful contextual parameters to tag ambiguous and unknown words. Our tagger is mainly supported by word lists and one corpus with around 10{sup}4 words. The system has been tested for texts of the so called "news" genre and is still on continuous development.