토픽모델링과 키워드 네트워크 분석을 활용한 국내 가짜뉴스 연구 동향 분석

Abstract

Purpose – This study analyzes the research trends of fake news in South Korea from 2017 to 2024 using advanced text mining techniques. Design/Methodology/Approach – A total of 474 abstracts were collected from the DBPIA academic database, and these were analyzed through keyword frequency analysis, Latent Dirichlet Allocation (LDA) topic modeling, and keyword co-occurrence network analysis. The data was divided into three periods, 2017–2019, 2020–2021, and 2022–2024, to track the temporal evolution of research topics. Findings – In the early phase (2017–2019), research mainly focused on the spread of fake news via social media and the nature and impact of misinformation. During the next phase (2020–2021), as the COVID-19 pandemic unfolded, research shifted to the spread of false information related to public health, particularly concerning the virus and its global implications. In the most recent period (2022–2024), research expanded to cover a wide range of issues, including the intersection of fake news and national security, as well as the role of digital platforms in the rapid dissemination of misinformation. Research Implications – This study highlights how fake news research has evolved in response to both social and technological changes. It underscores the shift in academic focus from social media dynamics to broader concerns related to security and digital technologies. Additionally, it lays the groundwork for future research, particularly addressing challenges posed by generative AI and the growing influence of digital platforms.