Derivation of Macedonian verbal adjectives
Journal
Proceedings of international conference" recent advances in natural language processing"(RANLP'07)
Date Issued
2007-09-27
Author(s)
Petrovski, Aleksandar
Abstract
Macedonian printed lexicons contain very few verbal
adjectives, which results in regular collapsing of spelling
checkers. In order to overcome this problem, we
automatically derived the most common verbal adjectives.
This paper describes the derivational process, which resulted
in more than 30000 new adjectival lemmas derived from
nearly 20000 verbs originating from printed dictionaries.
Generation of all the inflections of derived adjectives is also
presented. Both processes were performed with the linguistic
development environment INTEX/NooJ. Accuracy of the
obtained adjectival lemmas and their inflective forms was
tested on the lexicon containing the most frequent word forms
appearing in Macedonian daily newspapers and in recently
published books. Even with this limited validation lexicon,
many newly derived adjectives have been recognised in
standard language.
adjectives, which results in regular collapsing of spelling
checkers. In order to overcome this problem, we
automatically derived the most common verbal adjectives.
This paper describes the derivational process, which resulted
in more than 30000 new adjectival lemmas derived from
nearly 20000 verbs originating from printed dictionaries.
Generation of all the inflections of derived adjectives is also
presented. Both processes were performed with the linguistic
development environment INTEX/NooJ. Accuracy of the
obtained adjectival lemmas and their inflective forms was
tested on the lexicon containing the most frequent word forms
appearing in Macedonian daily newspapers and in recently
published books. Even with this limited validation lexicon,
many newly derived adjectives have been recognised in
standard language.
Subjects
File(s)![Thumbnail Image]()
Loading...
Name
RANLP2007.pdf
Size
26.51 MB
Format
Adobe PDF
Checksum
(MD5):cc1d95805336be6f73acd52f7bb523fd