Thinking beyond the nouns - computing semantic relatedness across parts of speech
Iryna Gurevych · TUbilio (Technical University of Darmstadt) · 2006
Semantic relatedness is any kind of lexical-functional association between two words. For example, given the words “surfen”, “Internet” and “Apfel”, human judges are likely to classify “surfen Internet” as being more closely related than “Internet Apfel” or “surfen Apfel”. Information about semantic relatedness of words is often required in various natural language processing (NLP) applications ranging from spelling correction (Budanitsky & Hirst, 2005) and sense tagging (McCarthy et al., 2004) to information retrieval (Gurevych 2005a).