Mathematical models of systems of transcription of name-and-surname groups

In processing foreign personal names and surnames, the task of finding how to represent the way they sound in Russian often arises. This is important, for example, for intellectual property and other property paperwork, for translations, and for creating news stories. Information is often not available about the language of representation of the personal name. Therefore, the problem of automated identification of the language by the structure of the name-and-surname group is actual. Another problem is the automation of rule creation processes for the transcription of personal names. This paper proposes an algorithmic approach to solve these problems using the methods of machine learning.