On certain aspects of Kazakh part-of-speech tagging

Aibek Makazhanov, Zhandos Yessenbayev, Islam Sabyrgaliyev, Anuar Sharafudinov, Olzhas Makhambetov

2014

Closed access Metadata (CSL-JSON)
DOI 10.1109/icaict.2014.7035953
Data confidence 70 medium
How this score is calculated

It is built from checkable properties of the record itself: whether the article has a DOI, how many independent sources confirmed its data, whether the required fields are filled in, and how recent the latest confirmation is. Disagreement between sources and a missing identifier lower it. A value of 80 or above means the record was published without further checking; 60 to 79 means it is published at reduced confidence, with a manual check advisable. A record verified by an editor by hand is never pushed below 90 by a recalculation. This scores our data about the article, not the article itself.

Abstract

We compare and discuss various approaches to the problem of part of speech (POS) tagging of texts written in Kazakh, an agglutinative and highly inflectional Turkic language. In Kazakh a single root may produce hundreds of word forms, and it is difficult, if at all possible, to label enough training data to account for a vast set of all possible word forms in the language. Thus, current state of the art statistical POS taggers may not be as effective for Kazakh as for morphologically less complex languages, e.g. English. Also the choice of a POS tag set may influence the informativeness and the accuracy of tagging.