Tracing the Pace of COVID-19 Research: Topic Modeling and Evolution
Published in Big Data Research, 2021
Keywords: COVID-19, Deep learning, Topic modeling, Bibliometric analysis, Science of Science
COVID-19 has been spreading rapidly around the world. With the growing attention on the deadly pandemic, discussions and research on COVID-19 are rapidly increasing to exchange latest findings with the hope to accelerate the pace of finding a cure. As a branch of information technology, artificial intelligence (AI) has greatly expedited the development of human society. In this paper, we investigate and visualize the on-going advancements of early scientific research on COVID-19 from the perspective of AI. By adopting the Latent Dirichlet Allocation (LDA) model, this paper allocates the research articles into 50 key research topics pertinent to COVID-19 according to their abstracts. We present an overview of early studies of the COVID-19 crisis at different scales including referencing/citation behavior, topic variation and their inner interactions. We also identify innovative papers that are regarded as the cornerstones in the development of COVID-19 research. The results unveil the focus of scientific research, thereby giving deep insights into how the academic society contributes to combating the COVID-19 pandemic.
Citation:
@article{liu2021tracing,
title={Tracing the Pace of COVID-19 Research: Topic Modeling and Evolution},
author={Liu, Jiaying and Nie, Hansong and Li, Shihao and Chen, Xiangtai and Cao, Huazhu and Ren, Jing and Lee, Ivan and Xia, Feng},
journal={Big Data Research},
volume={25},
pages={100236},
year={2021},
publisher={Elsevier}
}