Exploratory Analysis of Scholarly Publications on Artificial Intelligence (AI) in Colonoscopy using Litstudy

Mary Adewunmi

Keywords: Litstudy, Artificial Intelligence, Colonoscopy, Non-Negative Matrix Factorization(NMF), Natural language Processing(NLP)

TL;DR: Using Litstudy, we conducted an exploratory bibliographic data analysis of scholarly publications using artificial intelligence (AI) in colonoscopy as a usecase.

Abstract:

Due to the large number of scholarly papers on AI and colonoscopy and the short research period, it can be difficult to answer general questions about the research area, such as who the key authors are and what the key issues or insights are. We use Litstudy, a Python library, to study colonoscopy AI research. ”AI” and ”colonoscopy” keywords were used as search results. 3865 IEEE Xplore and
2007 Springer bibliographies were downloaded. Scopus found 5083 citation papers, excluding 789 unavailable citations. Topic clusters were created using the NMF model with a 0.85 threshold. Topic clouds showed that ”Patient” occurred most in four topics: 2, 3, 7, and 10. Despite querying IEEE, Springer, and Scopus databases with the ”Artificial Intelligence” keyword, subject 5 with AI has the lowest topic phrase weight in topic clouds. Topic 10 words cluster on colon cancer rehabilitation in colonoscopy showed weak topic clusters. The project selects scientific articles, analyses and visualises their scholarly contribution using natural language processing (NLP), bibliographic network analysis, and, most importantly, reveals word clusters in AI for colonoscopy publications.