polysemysimilarity
Polysemysimilarity is a proposed metric in linguistics and natural language processing that quantifies how similar the different senses of a single polysemous word are to one another. It focuses on the internal relatedness of a lemma’s senses rather than the similarity between different lemmas. The concept is used to analyze lexical complexity, inform sense disambiguation, and support semantic search and lexicography.
Approaches to measuring polysemysimilarity fall into three broad families: knowledge-based, distributional, and context-driven. Knowledge-based methods use
Data sources include sense-annotated corpora and lexical databases; applications include disambiguation, machine translation alignment, information retrieval
Challenges include defining appropriate sense granularity, cross-language applicability, and domain shift. The term polysemysimilarity is not