Using complex networks to identify patterns in specialty mathematical language: a new approach

The special words and combinations of two words (so-called collocations) that are specific references within a specialty language collectively constitute the terminology of that specialty language and characterize that language. In this paper, we analyze the specialty mathematical language produced by the scientific community of complex networks from the perspective of the theory and tools of complex networks to extract new properties and characteristics of this language. For this purpose, a linguistic corpus of 86 extended abstracts of papers (all of them based on the theory and applications of complex networks) has been used. To analyze the collocations that are part of the language terminology, a multilayer network and the line graph structure have been considered. The obtained results of our computational study show that the structure of collocations in this mathematical specialty language text corpus referred to complex networks is much more sensitive to variations in the intensity of their appearance in specialized texts than the structure of the words isolated from the specialty text itself, pointing to an intrinsic feature in the linguistic structure of this type of texts. This model can be a very useful tool for translators who work with specialized texts in the area of knowledge corresponding to the corpus studied, a particularly difficult task if the translator is not a specialist in this field.