텍스트 마이닝과 네트워크 분석을 활용한 출판-AI 연구 동향 탐색

초록

As artificial intelligence technology advances rapidly, academic research on publishing-AI has been expanding both quantitatively and qualitatively. This study applied TF/TF-IDF analysis, network centrality analysis, and CONCOR analysis to academic literature in the publishing-AI domain to identify the distribution of core conceptual terms, the structural characteristics of the research network, and latent thematic clusters. Results showed that ‘education’ recorded he highest TF frequency, while ‘education’, ‘French’, ‘translation’, and ‘technology’ achieved high TF-IDF specificity values, confirming their status as specialized core topics intensively discussed within particular research clusters. Network centrality analysis identified ‘technology’ as the central hub node with the highest degree centrality. CONCOR analysis further revealed four semantic clusters: technological transformation of AI-based digital publishing (Cluster 1), innovation in AI-based educational and learning content (Cluster 2), AI-based multilingual and literary content production (Cluster 3), and restructuring of linguistic and lexical knowledge systems driven by AI adoption (Cluster 4). These findings demonstrate that publishing-AI research forms a multilayered landscape in which education, language, and culture are organically interconnected around technological and industrial transformation, providing foundational data for future research directions and industrial strategy in the publishing-AI domain.

키워드