Automaatne ajaväljendite tuvastamine eestikeelsetes tekstides
Automatic recognition and normalization of temporal expressions in Estonian language texts
Author(s): Siim OrasmaaSubject(s): Language and Literature Studies
Published by: Eesti Rakenduslingvistika Ühing (ERÜ)
Keywords: computational linguistics; semantic markup; annotation; temporal expressions; Estonian
Summary/Abstract: In this article, I give an overview of the task of temporal expression recognition and normalization in natural language texts and describe a rule-based system, which was designed to solve this task in Estonian language texts.A TimeML-based markup language was used to annotate temporal expressions. Following the markup possibilities in TimeML, I define two rough criteria for deciding whether a temporal expression is markable or not: 1) a markable expression should contain temporal information of year, month, week, day, hour or minute granularity, or 2) a markable expression should contain a reference to the past, present or future, anchored to the time of creation of a document or to some other markable expression.
Journal: Eesti Rakenduslingvistika Ühingu aastaraamat
- Issue Year: 2012
- Issue No: 8
- Page Range: 153-169
- Page Count: 16
- Language: Estonian