Currently, the role of scientific research achievements in policy-making is becoming increasingly prominent. However, the implicit relationships between them are still unclear and difficult to uncover, making it urgent to build systematic tools to help policymakers accurately and effectively select scientific evidence. From the perspective of policy citation, this study aims to construct a policy and paper association graph based on large-scale data, and on this basis, provide scientific paper recommendations for policy-making. In the graph construction phase, using Overton and OpenAlex databases as the primary data sources, a knowledge graph schema layer integrating multiple types and hierarchical entities (including 9 types of entities and 5 types of relationships) is designed and constructed. A top-down knowledge graph construction strategy is adopted to extract and generate 47,327,880 semantic triples, which are stored in a Neo4j graph database to enable efficient querying and visualization support. In the graph application phase, six knowledge graph inference techniques are used, and the results are evaluated with five evaluation metrics. The results indicate that the proposed method can recommend scientific papers for policy-making more accurately and efficiently. The research not only provides a feasible framework and specific suggestions for achieving government decision-making based on knowledge graph technology but also offers valuable theoretical and practical foundations for exploring intelligent policy support tools.
关键词
引文分析/知识图谱/政策引用论文/论文推荐/循证决策
Key words
Citation Analysis/Knowledge Graph/Policy Citation of Papers/Paper Recommendation/Evidence-Based Decision-Making
评论