This paper presents a semi-automated approach to extract and name research topics from publicly funded research. The authors applied this method to $1.9B of National Cancer Institute (NCI) funding over 21 years in the radiological sciences. The aim was to identify micro- and macro-scale research topics and funding trends. The method relies on sequential clustering of existing biomedical-based word embeddings, naming using subject matter experts, and visualization to discover trends at a macroscopic scale above individual topics. The results revealed two dominant axes: physics-biology and therapeutic-diagnostic. The authors found that funding for therapeutics- and physics-based research have outpaced diagnostics- and biology-based research. The findings could provide insights to funders on their funding allocation, assist investigators in contextualizing their work, and allow the public to review where their tax dollars are being allocated.
Publication date: June 23, 2023
Project Page: N/A
Paper: https://arxiv.org/ftp/arxiv/papers/2306/2306.13075.pdf
