Thematic Clusters within Text Sets and Examining Temporal Dynamics of Themes (Using Conference Proceedings in "Argument Mining")
This paper presents the results of examining changes that occur in thematic clusters constructed on a text set of conference proceedings from the research field "Argument mining". Identification of terms, analysis of their relations and thematic clustering are performed with the use of a third-party software that allows to extract terms in the form of noun phrases and to cluster them in accordance to a modularity based algorithm. The quality of the resulting clusters is estimated by employing three distinct criteria. Temporal transformation of terminological content of clusters is analyzed through the use of directed graphs constructed with an underlying criterion that enables recognition of the most significant changes. Terminological lexicon of the identified thematic clusters characterizes distinct directions in which studies are conducted, while transformation of their terminological content demonstrates shifts of researchers' interests.