AI 기반 텍스트 마이닝을 통한 반부패 연구 동향 분석: 1995–2025년 한국부패학회 논문 LDA 모델링을 중심으로 = Trend Analysis of Anti-Corruption Research via AI-Based Text Mining: LDA Modeling of Korean Association for Corruption Studies Papers, 1995–2025

다국어 초록 (Multilingual Abstract)

This study employs Latent Dirichlet Allocation (LDA) topic modeling to examine the trends and intellectual structures of anti-corruption research in South Korea, concentrating on 424 Korean- language abstracts and 516 English-language abstracts derived from a total of 625 articles published in the Korean Corruption Studies Review over the last 30 years (1995–2025). This work fills a significant research gap by utilizing large-scale computational text analysis, in contrast to prior studies on anti-corruption scholarship in Korea, which predominantly utilized qualitative methods focused on institutional analysis and case studies. This study presents three research questions. What are the principal research themes delineated in the journal from 1995 to 2025? What transformations and reorganizations have these subjects undergone throughout history? What implications do LDA analytical results present for forthcoming anti-corruption policy and scholarly discussion? The findings indicate three primary themes in the Korean abstracts: the prevention of corruption and administrative trust in public organizations, the protection of whistle-blowers and legal-institutional responses, and the social impact of corruption and policy responses, reflecting a balanced research trend across institutional trust, organizational ethics, and policy responsiveness. Conversely, eight unique issues were identified in the English abstracts, encompassing worldwide anti-corruption policy, internal audit and integrity enhancement, whistle-blower protection, governmental innovation, and the Fourth Industrial Revolution. This study illustrates the temporal fluctuations in the importance of research topics as influenced by evolving political and social situations. This study offers a basic reference for future policy design and scholarly discourse in anti-corruption research by reconstructing its topic evolution through computational text mining.