Research for indexing, editing, and retrieving mathematical information
In order to enhance our information services for the mathematical community, we develop innovative tools for editing, indexing, and retrieving mathematical information in joint research projects with our partners. In addition to the scientific terminology, mathematics has developed its own language, i.e., mathematical formulae that provide complex content in very precise, condensed form.
Our research focuses on the following topics:
- Analyzing and developing automatic methods and processes for semantically indexing and enriching data and documents. Establishing a controlled, discipline-specific vocabulary, including relations to other multi-lingual ontologies (e.g., the MSC2010 version based on SKOS (Simple Knowledge Organization System) and the Semantic Multilingual Glossary for Mathematics (SMGloM).
- Developing heuristic methods to establish links between data and documents on the one hand, and between these data and documents and third-party sources on the other hand. These methods are used, for example, to compare bibliographical references with documents referenced in zbMATH by means of semantic analysis. Similar methods are used to link from papers published in journals to the corresponding preprints in arXiv, or to match zbMATH articles containing references to mathematical software against records in the software database swMATH. These methods are to be expanded and further developed in order to include more resources and research data from third parties.
- Developing methods for automatic disambiguation of author information. To this end, information on research areas, co-authors, external identifiers, publication periods, and biographical data are analyzed. In addition, these methods are to be further developed so that they can be used to standardize the names of institutions.
- Further developing the integrated formula search by enriching data with semantic information. The formula search allows for a very precise search for mathematical formulas or formula schemas – either independent of or combined with a traditional text search.