Automatisk identifiering av semantisk föränd - ring med hjälp av lexikala distributionella mönster

Abstract

By looking at word relations in corpora from different time periods we aim to automatically find candidates of semantic variation. We start with the verbal predicate, nominal object relation. This is a syntactic relation that also has an obvious semantic relation. By using techniques from collocational analysis and other statistical measures, we convey differences in distribution that can be indications of semantic variation.