Towards a recommender system for the choice of UDC code for mathematical articles

Abstract

Authors of scientific papers in the field of mathematics usually use the universal decimal classification scheme to search for related articles. UDC is a hierarchical classification scheme that allows librarians and editors to specify one or more codes for publications. Typically, the classification code identifies a subject editor who is responsible for the review process for articles submitted to scientific journals. In this article, we will explore a new approach to assigning UDC code for mathematical work, based on the OntoMathPRO ontology. This ontology is an applied ontology for the automatic processing of professional mathematical articles in Russian and English. An ontology defines concepts commonly used in mathematics, as well as an evolving and poorly established vocabulary extracted from contemporary scientific articles. OntoMathPRO covers a wide range of areas of mathematics such as number theory, set theory, algebra, analysis, geometry, computation theory, differential equations, numerical analysis, probability theory, and statistics. Each class has a textual explanation, Russian and English inscriptions, including synonyms. We investigated a set of classification functions, which are presented as ontology concepts, and identified the most relevant ones for constructing code maps of some UDC codes in the field of mathematics. We found that the code maps of the considered UDC codes can be built on the basis of the selected features (method, equation, problem). The values of these features are determined using the OntoMathPRO ontology. The constructed code maps allow for successfully assigning the considered UDC codes for publications.