On the application of extractive summarization methods for preparing scientific text abstracts

This study provides an overview of automatic summarization of scientific texts as a means of generating concise versions that retain key sentences and essential information from the original document. Automatic processing of scholarly texts has the potential to substantially streamline the work of researchers by facilitating the extraction of core information from scientific publications. The paper outlines the main components of scientific texts and the typical structure of research articles, reviews existing approaches to automatic text summarization, and discusses the most widely applied extractive methods, namely TF-IDF and TextRank. These methods employ different strategies for information extraction, taking into account structural relationships between sentences and the selection of keywords. However, their effective application to scientific texts may require further refinement to address domain- specific characteristics.