Structuring the topics of risk-related articles is expected to lead to understanding the characteristics of various risk sciences. Here, we used structural topic modeling to analyze the topics of risk-related articles, and compared each topic with the content of the encyclopedia of risk research. The target journals were Risk Analysis (RA), Journal of Risk Research (JRR), and Science published up to 2020. For RA and JRR, all articles were included, and for Science, articles that included risk in the title, abstract, or keywords were selected. 3761 RA articles, 1403 JRR articles, and 947 Science articles were included. Articles were classified based on the co-occurrence of terms used in the abstract. The 23 classified topics can be divided into three types of topics: risk events, risk analysis approaches, and risk sociality and communication. RA and JRR contained more articles on topics related to risk analysis approaches and risk sociality and communication in general, while Science included more articles on topics such as “diseases and genes.” In comparison with the encyclopedia, chapters 1–4 of the encyclopedia corresponded to topics on risk analysis approaches and risk sociality and communication, and chapters 6–9 corresponded to topics on risk events. The encyclopedia also had other chapters on finance and other topics. On the other hand, there was no chapter in the encyclopedia corresponding to the topic of “diseases and genes.” Analyzing the risk science by structural topic modeling can provide useful knowledge for overviewing the risk science fields and preparing the next encyclopedia.
View full abstract