토픽 모델링을 활용한 공공부문 인공지능 기반 지식관리 연구의 동향 분석 = A Trend Analysis of AI-Based Knowledge Management Research in the Public Sector Using Topic Modeling

다국어 초록 (Multilingual Abstract)

This study examines core themes and trends in artificial intelligence (AI)–based knowledge management research in the public sector and proposes future research directions and policy implications. Using topic modeling techniques, it analyzes academic publications from the past decade indexed in the Web of Science database. The results show a steady growth in related publications and a sharp increase in citations since 2020, reflecting rising scholarly interest. Key themes include learning, knowledge, networks, transfer, data, neural networks, graphs, deep learning, and discovery, highlighting strong links to AI, machine learning, and deep learning technologies. Latent Dirichlet Allocation (LDA) analysis identified five major topics: datasets, knowledge transfer, machine learning, deep learning–based training, and knowledge-based systems. These topics form two broad clusters: one focused on technical approaches such as transfer learning, cross-domain learning, and anomaly detection, and the other emphasizing application-oriented areas including image processing, object recognition, human perception, and signal processing. In addition, Latent Semantic Indexing (LSI) analysis captured recent AI research trends such as graph theory, domain adaptation, natural language processing, knowledge graphs, and health data analytics. Based on these findings, the study suggests strategic directions for developing effective AI-driven knowledge management frameworks in the public sector.