A Finite-State Morphological Analyser for Evenki

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 2581–2589 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 2581 A Finite-State Morphological Analyser for Evenki Anna Zueva, Anastasia Kuznetsova, Francis M. Tyers School of Linguistics, Department of Linguistics, Department of Linguistics Higher School of Economics, Indiana University, Indiana University anna.zueva.v@gmail.com, anakuzne@iu.edu, ftyers@iu.edu Abstract It has been widely admitted that morphological analysis is an important step in automated text processing for morphologically rich languages. Evenki is a language with rich morphology, therefore a morphological analyser is highly desirable for processing Evenki texts and developing applications for Evenki. Although two morphological analysers for Evenki have already been developed, they are able to analyse less than a half of the available Evenki corpora. The aim of this paper is to create a new morphological analyser for Evenki. It is implemented using the Helsinki Finite-State Transducer toolkit (HFST). The lexc formalism is used to specify the morphotactic rules, which define the valid orderings of morphemes in a word. Morphophonological alternations and orthographic rules are described using the twol formalism. The lexicon is extracted from available machine-readable dictionaries. Since a part of the corpora belongs to texts in Evenki dialects, a version of the analyser with relaxed rules is developed for processing dialectal features. We evaluate the analyser on available Evenki corpora and estimate precision, recall, and F-score. We obtain coverage scores of between 61% and 87% on the available Evenki corpora. Keywords: evenki, fst, morphology 1. Introduction Morphological analysis is an essential part of natural lan- guage processing (NLP), especially for languages with rich morphology. Morphological analysers detect the structure of a word form and return a lemma and corresponding gram- matical tags. Apart from morphological analysers, there are morphological generators, which use lemmas and grammat- ical tags to produce a particular word form. Morphological analysers and generators are widely implemented in various NLP applications such as corpus annotation, information re- trieval, machine translation, speech recognition, speech syn- thesis, and proofing tools. Evenki is an endangered Tungusic language spoken in Rus- sia (see Figure 1), China and Mongolia. According to Ethno- logue it currently has 15,800 speakers in total1. Since Evenki has rich morphology, a morphological analyser is extremely useful for processing Evenki texts. Without a morphological module, less than one third of the tokens in Evenki texts can be found in dictionaries and therefore can get part-of-speech tag or translation. Table 1 shows the comparison2 between the number of tokens in Evenki and English (as a language with little inflection), which can get analysis without morpho- logical processing. Although two morphological analysers for Evenki have al- ready been developed (see Section 2.2.), they are only able to produce analyses for at most 58% of Evenki corpora in the literary language3 and 44% of dialectal Evenki corpora. The aim of this work is to create a new morphological anal- 1https://www.ethnologue.com/language/evn (retrieved 2020-03-06). 2Evenki corpora and Evenki dictionaries are described in Sec- tion 2.1. and Section …

Content truncated to 500 words