Tagging Icelandic Text using a Linguistic and a Statistical Tagger

Hrafn Loftsson

Abstract: We describe our linguistic rule-based tagger IceTagger, and compare its tagging accuracy to the TnT tagger, a state-of-the-art statistical tagger, when tagging Icelandic, a morphologically complex language. Evaluation shows that the average tagging accuracy is 91.54% and 90.44%, obtained by IceTagger and TnT, respectively. When tag profile gaps in the lexicon, used by the TnT tagger, are filled with tags produced by our morphological analyser IceMorphy, TnT's tagging accuracy increases to 91.18%.