Identification of actual energy policy and security research topics in Elsevier and MDPI publications for 2021–2023. Part 2. Clustering of publications. Automatic extractive referencing

Abstract

This article focuses on identifying relevant research topics in Elsevier and MDPI publications 2021–2023 on energy policy and security based on publication clustering. It is shown that name groups extracted from titles and abstracts can be used to cluster publications to identify relevant research topics. The extraction of noun phrases was carried out using the Sifaka software package. A table of the 5,000 most frequent name groups was constructed, the terms of which were retained in the title and abstract texts, and the remaining terms were deleted. The prepared texts were used to cluster the bibliometric records using a demo version of Carrot2 and the Lingo3G algorithm. The annotation texts for each cluster were subjected to automatic extractive referencing using the sumy program with the Text-rank algorithm and the stopwords-en-ISO.txt list. The possibility of using extractive referencing technology to describe the dominant topic of individual publication clusters is demonstrated. A list of dominant themes in energy policy and security publications in Elsevier and MDPI publications for 2021–2023 is given.