Indexing mathematical scholarly papers as linked open data

Abstract

We present our work on developing an open source software platform for mining Linked Open Data (LOD) representation for a given collection of mathematical scholarly papers. Currently, the LOD cloud lacks up-to-date data on professional level mathematics. The main reason behind this is due to practical difficulties arising while dealing with such severe documents for indexing as mathematical papers that abound with formulas and specific structural elements ignored by the most state-of-the-art academic search engines. Our proof of concept demonstrates a feasible approach to parse these documents properly, dissect the semantics of their significant parts with the help of the ad hoc math-aware vocabulary, and publish their contents and metadata as RDF data. The authors argue that the platform at the final stage of its development cycle may be helpful for modern online scientific collections. For our experimental setup, we choose Math-Net.Ru – a digital collection well-known in the Russian mathematical community.