Topic Modeling Analytics of Digital Economy Research: Trends and Insights

This paper incorporates scholarly articles, conference papers, and published books, analyzing these sources to explore how topic modeling is used to uncover trends, identify research areas, and contribute to understanding the digital economy. The study utilizes bibliography data and Python libraries for topic modeling to examine 8,321 documents from the Scopus database. Through rigorous analysis, three distinct topics were identified, and their development trends were traced, providing insights into the shifting focus within the field. The study also validates the classification proposed by the coherence coefficient and contributes theoretically and methodologically by employing Latent Dirichlet Allocation (LDA) topic modeling technology within text analytics. The results of this study indicated that “Digital Transformation” emerged as the most popular, accounting for 56.6% of tokens. However, the “Digitalization” topic exhibited relatively lower popularity, representing only 21.2% of tokens. The findings help enhance the understanding of global research trends and offer a valuable framework for comprehending digital economy research. Furthermore, the study emphasizes the significance of analyzing digitalization, data governance, and digital transformation, highlighting the efficacy of LDA as a powerful tool for efficient and accurate text analytics. The research findings are especially pertinent for scholars in information science, data mining, econometrics, and bibliometrics. They offer a foundational understanding of topic modeling, facilitating further investigation and exploration in these fields. However, the study identifies two limitations, including the limited dataset extracted solely from Scopus and the restriction to abstracts rather than full texts. Future research should consider expanding data sources and incorporating full texts from authoritative articles in multiple languages to attain more comprehensive outcomes.