Tagging Icelandic text: A linguistic rule-based approach

Abstract

The Icelandic language is a morphologically complex language, for which a large tagset has been created. This paper describes the design of a linguistic rule-based system for part-of-speech tagging Icelandic text. The system contains two main components: a disambiguator, IceTagger, and an unknown word guesser, IceMorphy. IceTagger uses a small number of local elimination rules along with a global heuristics component. The heuristics guess the functional roles of the words in a sentence, mark prepositional phrases, and use the acquired knowledge to force feature agreement where appropriate. IceMorphy is used for guessing the tag profile for unknown words and for automatically filling tag profile gaps in the lexicon. Evaluation shows that IceTagger achieves 91.54% accuracy, a substantial improvement on the highest accuracy, 90.44%, obtained using three state-of-the-art data-driven taggers. Furthermore, the accuracy increases to 92.95% by using IceTagger along with two data-driven taggers in a simple voting scheme. The development time of the tagging system was only seven man-months, which can be considered a short development period for a linguistic rule-based system.

Augmenting a BiLSTM Tagger with a Morphological Lexicon and a Lexical Category Identification Step

“…A comparison of our model to other previously published taggers for Icelandic is shown in Table 2. The results for TnT, IceTagger and Ice-Stagger are presented in (Loftsson et al, 2009;Loftsson, 2008;Loftsson and Östling, 2013), respectively. All the reported results are fully comparable as they are based on exactly the same cross-validation split of the IFD corpus, with the exception that the TnT tagger does not employ data from DMII, and has therefore a higher ratio of unknown words.…”

Assessment of language skills weaknesses and their confrontation through personalized e-lessons

“…The disambiguation process will be tested using the latest methods of the current research on tagging algorithms, i.e. rule-based [7,9,14,15], stochastic [3,4,5,10,11] and transformation-based [1] algorithms. In order to achieve the best results regarding the morphosyntactic properties of Modern Greek, the most suitable tagger will be selected.…”

Tagging Icelandic text: an experiment with integrations and combinations of taggers

“…For each word class there is a predefined number of additional characters (at most six) which describe morphological features, like gender, number and case for nouns; degree and declension for adjectives; voice, mood and tense for verbs, etc. The reader is referred to (Loftsson, 2006a;Pind et al, 1991) for a more complete description of the tagset.…”

“…In this section, we describe four integration methods, all of which have resulted in an improved tagging accuracy of Icelandic text. The first two methods, which consist of integrating our morphological analyser with state-of-the-art DDT, are described in more detail in (Loftsson, 2006a). The latter two methods are new.…”

“…In (Loftsson, 2006a), we had improved the first simple voting result for Icelandic text by combining TBL*, TnT* and Ice-obtaining an accuracy of 92.94% (see row 4 in Table 3). Here, we improve this result by adding the taggers MXP and MBT to the combination pool, resulting in an accuracy increase to 93.29% (see row 5).…”