Text-mining Analysis of Health related to climate change in the abstracts of PUBMED from 1990 to 2019

Abstract

BACKGROUND AND AIM: This study attempts to understand the overall research trends in health related to climate change using a text mining method, which aims at extracting and processing meaningful information from vast texts through natural language processing and data mining techniques from structured or unstructured text data. Climate change, such as global warming and extreme weather events, has become a reality around the world and is expected to continue for a considerable period of time. The effects of climate change on health and disease have been suggested. METHODS: This study used the following keywords to search for articles of health related climate change published in 1990 to 2019 in PubMed: climate change, health, disease. We established a term-document matrix which describes the frequency of terms that occur in documents after removing special characters, numbers and stop-words. To investigate systematically the research trend of climate change studies, we made bigram from the term-document matrix and conducted frequency analysis and LDA(Latent Dirichlet Allocation) topic modeling. RESULTS:The keywords as global warming, public health, human health, infectious disease, vector borne disease, and greenhouse gas were ranked at the top with some ranking variations from 1990 to 2019. Particulate matter and sustainable development have increased in frequency since 2015. In terms of disease, the keywords of high frequency changed in the order of skin cancer, respiratory disease, cardiovascular disease, and mental illness within 1990 to 1999. CONCLUSIONS:We could see that the field of study of climate change and health with the times and including researches in 2020, new emerging infectious diseases and pandemics are expected to emerge as new keywords. KEYWORDS: Climate change, systemic review, text mining analysis, topic modeling, research trend