High-density knowledge rich contexts

Resumen

Knowledge Rich Contexts (KRCs) are one of the usual data categories contained in terminological
knowledge bases. In this paper we show how to extract KRCs that combine various items of knowledge to
facilitate the work of terminographers as well as user knowledge acquisition. For this reason, a new
knowledge pattern-based sketch grammar was designed within the corpus analysis tool Sketch Engine and applied to the EcoLexicon Environmental Corpus. After compiling the corpus with the sketch grammar, KRCs can be extracted through customized semantic word sketches (WS), which provide access to concordances where the item queried is related to others through one or several semantic relations. Then new queries are performed, reusing these new WS in order to collect high-density KRCs, which combine different semantic relations in a single sentence. In this paper we provide a characterization of high-density KRCs based on the amount of knowledge, concept types, conceptual depth, and number and type of conceptual relations codified in them.