ABSTRACT
OBJECTIVE: Data mining extracts meaningful information from large datasets. In this study, text mining techniques were used to extract keywords from the Journal of the Korean Society of Emergency Medicine, and the change trend was examined. METHODS: The rvest package in R was used to extract all papers published in the Journal of the Korean Society of Emergency Medicine from 2006 to 2016 that could be searched online. Among them, 3,952 keywords were extracted and studied. Using the selected keywords, the corpus was formed by refining keywords that did not correspond to MeSH (Medical Subject Headings) or were misspelled and had similar meanings based on agreement of researchers. Using the refined keywords, the frequencies of the keywords in the first and second halves of the studies were calculated and visualized. RESULTS: Word Cloud revealed that emergency medical service and cardiopulmonary resuscitation (CPR) were most frequently mentioned in both the first and second halves of the studies. In the first half, ultrasonography, stroke, poisoning, injury, and education were frequently mentioned, while in the second half, poisoning, injury, stroke, acute, and tomography were frequently mentioned. A pyramid graph revealed that the frequencies of emergency medical service and CPR were commonly high. CONCLUSION: Core keywords of the Journal of the Korean Society of Emergency Medicine were analyzed for correlations and trends. Changes in study topics according to key topics of interest and period were visually identified.