Topic Modeling of Scientific Publications in the Specialty of “Dentistry”

The intellectual analysis of texts and the task of identifying the thematic structure of documents allows us to solve various applied problems associated with documental retrieval, automatic selection of keywords and annotation of documents. In the parsing task of the dynamics of topics in a large number of scientific works, the automatic determination of text details of a work or article becomes fundamentally important. The use of machine learning methods and thematic modeling allows us to solve such problems. This paper analyzes the multiple data of the Russian Science Citation Index (dynamics of topics and keywords (phrases)) in the scientific specialty of Dentistry based on the analysis of more than 15,000 Russian texts of scientific articles published from 2005 to 2015. Thematic modeling of the corpus of available documents has been carried out on the base of the papers titles analysis. Using the methods of extracting keywords in the document and identifying the main topics in the corpus of documents, the dynamics and distribution of the top keywords and phrases, as well as the change of thematic focus of articles in the analyzed period have been analyzed.