Using Discourse Context to Interpret Object-Denoting Mathematical Expressions
Magdalena Anna Wolska, Mihai Grigore, Michael Kohlhase · Czech digital mathematics library · 2011
summary:We present a method for determining the context-dependent denotation of simple object-denoting mathematical expressions in mathematical documents. Our approach relies on estimating the similarity between the linguistic context within which the given expression occurs and a set of terms from a flat domain taxonomy of mathematical concepts; one of 7 head concepts dominating a set of terms with highest similarity score to the symbol’s context is assigned as the symbol’s interpretation. The taxonomy we used was constructed semi-automatically by combining structural and lexical information from the Cambridge Mathematics Thesaurus and the Mathematics Subject Classification. The context information taken into account in the statistical similarity calculation includes lexical features of the discourse immediately adjacent to the given expression as well as global discourse. In particular, as part of the latter we include the lexical context of structurally similar expressions throughout the document and that of the symbol’s declaration statement if one can be found in the document. Our approach has been evaluated on a gold standard manually annotated by experts, achieving 66% precision.