Abstract:
The paper presents an approach to discourse of science as a mental space whose structure displays the logic and tools of the scientific cognition and communication. We propose that this space is composed of the mental actions have been described previously [1], and a mental objects being described here as a definitions express the main ideas of an academic text. Based on more than 7000 research publications from peer-reviewed journals, an efficiency of the procedures of relationalsituational analysis and machine learning to reveal the structure of correlations between mental objects and mental actions is shown. It is also shown that in publications of various disciplines the correlations between mental objects and mental actions vary in composition and strength, and therefore this method can capture the specifics of the intellectual tools of research and communication in a particular discipline. A presenting of research publications content as a constellation of extracted mental actions and mental objects can be one of the way of the text mining enhancement.
Keywords:text mining, science discourse, research publication, mental actions, mental objects.
Fulltext:
PDF file (448 kB)
