engleski

How we color the world with words

This paper presents a computational approach to the automatic detection of language patterns, specifically those dealing with expressing colors in the Croatian language. It investigates different lexicalization patterns of color terms, mainly compounds and multiword units, in order to classify them and prepare them for usage in the design of an algorithm that will automatically recognize and annotate these expressions in Croatian text. The paper also presents a comparative analysis of different classes of color terms found in a corpus built from books intended for younger (CLC) and older (ALC) populations. Finally, the research data is presented through a dictionary of three types of color terms categorized as multiword expressions.

color terms ; lexicalization patterns ; multiword expressions ; natural language processing ; digital humanities ; Croatian ; NooJ

nije evidentirano