Document clustering and topic mdelling of scopus abstracts on phage therapy

In recent years, the resurgence of phage therapy has garnered significant attention as a promising alternative to traditional antibiotics, offering a potential solution to the growing threat posed by antibiotic resistance. This study presents an in-depth analysis of the evolving trends, key themes, and gaps in phage therapy research, employing advanced computational techniques to systematically review and categorise a vast dataset of Scopus abstracts. The collection initially comprised 6002 articles, which was refined to 5122 documents after filtering out those without abstracts. We then applied natural language processing (NLP) techniques to the document set, including various approaches for document clustering and topic modelling. The calculated coherence of the topic modelling, performed using the novel BERTopic approach, indicates a reasonable level of interpretability of the obtained topics. This study not only categorises the evolving trends, key themes, and gaps in phage therapy research but also showcases the potential of combining NLP and machine learning for organizing scientific literature. It provides a nuanced understanding of the study theme and highlights areas for further investigation and policy development. Although the study focussed on phage therapy research, which is of high relevance due to the potential of phage therapy as an alternative to traditional antibiotics, the methodology employed is versatile and can be applied to various other disciplines, including climate and agriculture, environmental sciences, and economics. These key findings and insights underscore the transformative potential of advanced computational methods in shaping the future of scientific research and policy-making, particularly in addressing the pressing challenge of antimicrobial resistance and guiding strategic initiatives across various sectors.