Scientific pertinence of developing machine learning technologies for the triage of COVID-19 patients: A bibliometric analysis via Scopus

Abstract

The COVID-19 pandemic poses challenges in terms of diagnosis, as existing tests often fail to consistently distinguish COVID-19 pneumonia from other acute respiratory infections. Although the integration of machine learning (ML) with diagnostic procedures holds promise in addressing this issue, a comprehensive understanding of the current scientific relevance and trends of this approach has not yet been undertaken. In this study, it was conducted a bibliometric analysis to explore the significance of developing ML-based algorithms for triaging COVID-19 patients. By querying the Scopus database and employing the Bibliometrix package in R software, a dataset was obtained and analyzed. VOSviewer software was chosen to construct bibliometric networks. From January 2020 to mid-2022, a total of 1309 articles across 572 journals have been published on ML-assisted COVID-19 diagnosis, treatments, and prognosis. Notably, Scientific Reports and Computers in Biology and Medicine have seen considerable contributions. The analysis involved 6086 institutions across ninety countries, with the United States and China being the most impactful. The Massachusetts General Hospital and Harvard Medical School emerge as key institutions, while Hao Chen from Westlake University (China) stands out among 10414 authors. Two major research trends were identified: the use of Random Forest ML algorithms with laboratory datasets for COVID-19 mortality prognosis, and the analysis of chest CT and X-ray images for early COVID-19 detection using Convolutional Neural Networks (Deep Learning). In summary, this bibliographic analysis discusses the rationale of selecting Scopus as the primary database, outlines a detailed methodology for evaluating article quality, addresses the limitations of co-citation network analysis, and explores potential variations in the adoption of different ML algorithms for COVID-19 diagnosis, treatment, and prognosis.

Ähnliche Arbeiten