Analyzing the Status of Endocrinology and Metabolic Research in Iran Using Text Mining Methods

Abstract

Introduction: Due to the importance and status of the endocrine field in the health sector, facilitating information retrieval in this field seems to be important. This study endeavored to run topic modeling of articles published by Iranian researchers in the field of endocrinology and metabolism in the science citation database. Methods: This descriptive study was done by text mining method. In this study, abstracts of articles were extracted from the science citation database using selected keywords of Medical Subject Headings (MeSH). 5552 articles were retrieved from 1977 to 2019, then the text of abstracts was analyzed and categorized in MATLAB software. Results: Subject categories with 20 items were extracted in 48 categories. Diabetes with 7145 recurrences has been considered by Iranian researchers more than other topics. Subject category related to metabolic syndrome diseases had the highest number of articles (304 articles) and subject category No. 47 which was related to osteoporosis had the lowest number of articles (51 articles). Conclusion: Iranian researchers have done more research on metabolic syndrome and less research on osteoporosis. Topic categories including Dwarfism, Parathyroid Diseases, Pituitary Diseases, Gonadal Disorders, Polyendocrinopathies, and Autoimmune that did not exist in the topics resulting from topic modeling indicate a gap in the research of Iranian researchers, that emphasizes the need for more attention to these areas.

Keywords

• topic modeling
• endocrinology
• metabolism
• text mining
• latent dirichlet allocation