Factores asociados a la citación de artículos biomédicos colombianos: análisis con Machine Learning

Abstract

Citation indicators can be used to measure the impact or usefulness of research results in a scientific article; however, this usage can be controversial. Intrinsic and extrinsic factors influence the citation of an article, not to mention that citation behavior can differ between thematic areas, which hinders the comparison between articles and disciplines. Understanding that context can affect citation analysis is essential to interpret indicators properly; for this reason, we want to recognize the factors that influence the citation of Colombian biomedical journals indexed in Scopus using Machine Learning algorithms. With ‘Gradient Boosting Classifier’ and ‘Light Gradient Boosting Machine’ algorithms, we find characteristics of importance such as the h-index of the first and last author, open access, number of authors and keywords of the article, in addition to identifying the number of pages. These characteristics are relevant to the area of interest and can provide context for future analyses, always considering that what should be relevant about an article is not how many citations it attracts but whether it helps to fill gaps in knowledge