Title :
Word sense disambiguation in Mongolian language
Author :
Bataa, Batzolboo ; Altangerel, Khuder
Author_Institution :
Department of Software Engineering, School of Computer Science and Management, Mongolian University of Science and Technology, Ulaanbaatar Mongolia
Abstract :
Word sense disambiguation is an important intermediate stage for many natural language processing applications, especially transformation from Cyrillic into Mongolian script. A word sense could be disambiguated by other words in the context as nouns, verbs used with the word. In this research, we have analyzed the result of an experiment on a word disambiguation system for Mongolian language based on statistical model, to which one sense per collocation algorithm is applied, and suggested a model established of the weight of sense rate and the weight of distance to the adjacent words to improve the accuracy. We chose one of 1.7 thousand words which have more than one sense that were used in the experiment, and performed an experiment on 41 thousand words. We researched one of 6.3 thousand words which have a complex stem that can be considered to have another stem plus a suffix structure. Since it is the first work in this field for Mongolian language no previous work results were available for comparison.
Keywords :
Bayes methods; classification; natural language processing; text analysis; Bayesian classifier; Cyrillic script; Mongolian language; Mongolian script; natural language processing application; noun; one sense per collocation algorithm; sense rate; statistical model; suffix structure; verb; word sense disambiguation; Accuracy; Context; Dictionaries; Educational institutions; Natural language processing; Training; Bayesian classifier; decision list; one sense per collocation; word sense disambiguation;
Conference_Titel :
Strategic Technology (IFOST), 2012 7th International Forum on
Conference_Location :
Tomsk
Print_ISBN :
978-1-4673-1772-6
DOI :
10.1109/IFOST.2012.6357625
