Abstract: Abstract: The force-directed layout method was a classical approach used for knowledge graph visualization, and it was employed to uniformly distribute nodes in the visualization space through attractive and repulsive forces between nodes. However, since visualization of domain-specific knowledge graphs was often closely related to domain applications, the classical force-directed layout method did not meet the visualization requirements of vertical-domain knowledge graphs. An improved force-directed layout method was proposed by analyzing the graph structure, data characteristics, and visualization requirements of query applications of medical record knowledge graphs, with the aim of supporting clinical decision-support applications. Based on the classical force-directed layout model, this method introduced attraction between nodes of the same level and same type, repulsion between nodes of the same level but different types, and spacing-preserving repulsion between non-leaf and non-adjacent nodes. As a result, nodes of the same level and same type were grouped into the same region, nodes of the same level but different types were distributed into separate regions, and nodes at different levels were arranged with reasonable spacing, thereby reducing node crowding and edge crossings. The final visualization layout clearly revealed the data’s type characteristics, quantity characteristics, and community characteristics. Additionally, node type and the number of same-level, same-type nodes were incorporated as factors determining edge length between nodes, so that node importance was expressed through edge length, further improving the rationality of the visualization layout and enhancing semantic expressiveness. Experimental results showed that, compared with the FR method, the proposed method achieved a 165.5% improvement in domain-specific semantic visualization capability. From the visualization results, better performance was obtained for electronic medical record knowledge graph visualization.

Key words: Keywords: medical record knowledge graph, visualization, force-directed layout, node link graph, force-directed method integrating type information

摘要： 力导向布局方法是知识图谱可视化采用的经典方法，其通过节点间的引力和斥力将节点均匀地分布在可视化空间内。但由于垂直领域的知识图谱可视化常与领域应用密切相关，因此，经典的力导向布局方法并不满足垂直领域知识图谱的可视化需求。该文章以支持临床辅助诊疗应用的病历知识图谱可视化技术为目的，通过分析病历知识图谱的图结构、数据特点以及查询应用的可视化需求，提出了改进的力导向布局方法。该方法在经典力导向布局模型的基础上引入同级同类型节点间引力、同级异类型节点间斥力和非叶且非邻接节点间维持间距排斥力，使得在布局上同级同类型节点能够集中在一个区域，同级不同类型节点分布到不同的区域，不同级别的节点以合理的间距分布，从而减少节点挤压和边交叉现象，使最终生成的可视化布局能够清晰地展示出数据的类型特征、数量特征和社区特征。方法还将节点的类型和同级同类型节点的数量作为决定节点间边长的因子，通过边的长度表达节点的重要性，从而进一步提升可视化布局的合理性，强化了语义的可视化表达。实验结果表明，本文的方法在特定领域语义可视化能力方面，相比FR方法提升了165 .5%。从可视化结果来看，该方法在电子病历知识图谱可视化上获得了更好的效果。

关键词: 关键词: 病历知识图谱, 可视化表达, 力导向布局, 节点链接图, 融合类型信息的力导向方法