Purpose: This study analyzes the topics of the agricultural products safety and quality R&D project conducted by the National Agricultural Products Quality Management Service and the Future Growth Forum for Agriculture, Forestry and Food. Using topic modeling, the study aims to compare the research priorities and policy trends of the two organizations and to identify differences in the topics and trends of agricultural and food R&D.Methods: The data included 942 project titles (1963–2025) and 173 policy forum topics (2011–2025). After text preprocessing (removing missing values, stop words, and merging compound nouns), word frequency analysis and Latent Dirichlet Allocation (LDA) were applied. To clarify the meaning of the topics, action-oriented nouns (e.g., research, analysis, investigation) were excluded from the secondary model, and the topic structure and semantic concentration were compared before and after exclusion.Results: In the analysis of agricultural products quality management, the most frequently mentioned terms were ‘investigation’, ‘research’, and ‘analysis’. However, excluding action-oriented nouns, more specific and semantically oriented keywords such as ‘data’, ‘inspection standards’, and ‘country of origin’ were emphasized. An integrated analysis of research and forum data yielded five themes: ‘residues’, ‘country of origin & discrimination’, ‘feed’, ‘country of origin & testing,’ and ‘agriculture & food’. While institutional research highlighted topics related to hazardous substance analysis and country of origin testing, the Future Growth Forum emphasized policy and future-oriented topics such as agricultural technology and future planning, suggesting a complementary yet distinct role between practical research and planning policies.Conclusion: This study utilized LDA-based text mining to analyze and visualize thematic trends and differences in roles across diverse actors in agricultural and food R&D. These results provide a foundation for strategic policy formulation and future research directions, demonstrating that topic modeling is an effective tool for research trend analysis and setting policy directions.
Kim et al. (2025) studied this question.