텍스트 마이닝을 통한 국내 미세플라스틱에 대한 비정형 데이터의 정보화

Objectives: In this study, we seek to provide a thorough insight into how people perceive them and uncover issues and hidden trends about the significant microplastic pollution problems by analyzing unstructured data on microplastics. Methods: Environmental news articles related to microplastics were collected. Text mining techniques including data pre-processing, word cloud, TF-IDF weight-based trend analysis, and topic modeling were used to analyze the amount of textual data. Results and Discussion: The public's interest in microplastics is consistently growing, according to an analysis of all environmental news and the keyword ‘microplastic’ from 2014 to 2021 conducted via BIGKinds. The keyword 'trash' was the overwhelmingly enormous weight among words. The top 5 keywords connected to microplastics did not fade away and continued appearing even though the socially noticeable keywords during the study period varied yearly. This indicates that the primary issue with microplastics related to keywords has not yet been solved. Our study has a limitation of subject diversity because we only focused on microplastic news. The results indicated that all processes from plastic pollution emergence to treatment, such as microplastic pollution sources, microplastic detection, and prevention methods against microplastics. Conclusions: Text mining analysis was performed on microplastics in environmental news. This study is meaningful in that it presents a new methodology for environmental and social problem analysis, and can be used as basic data for environmental policy establishment and problem-solving using unstructured big data.