Visualizationan alysis of domestic deep learning research trends based on co-word analysis method

This study employs the bibliometric co-word analysis method, taking 1,365 "deep learning"-related articles from the Peking University Core Journals and CSSCI-indexed academic journals in the CNKI database as research samples. Using COOC 14.9 and VOSviewer 1.6.19 tools, a dissimilarity matrix of high-frequency keywords was constructed for systematic clustering analysis. A keyword network hierarchy diagram was generated based on degree centrality weights to explore research hotspots. Additionally, a two-mode matrix integrating core authors and high-frequency keywords was developed to reveal the primary research domains of core authors and their interconnections. The results indicate that current deep learning research in China primarily focuses on five areas: theoretical research, blended teaching, artificial intelligence applications, deep teaching, and teaching reform. Based on the visualization analysis and mutation detection results, future research prospects in the field of deep learning are proposed, including technology integration, instructional design, empirical research, and teaching evaluation.