心力衰竭知识图谱的构建

Plotting knowledge graphs for heart failure

目的:利用现有资源构建专病知识图谱,为临床决策提供知识与技术支持。方法:以心力衰竭为例,首先明确疾病、检查、手术、药物、科室、症状、解剖学7类顶层语义概念以及2类属性和14类语义关系,其次复用现有词表、本体按照语义概念类型分别进行层级关系构建和同义术语归并,最后利用知识库资源进行语义关系抽取和同义术语归并。结果:形成包含524个概念、4 360对语义关系的心力衰竭知识图谱。结论:利用现有词表、本体和知识库资源采用知识融合的方法可以实现专病知识图谱的构建。

Objective To provide the knowledge and technologies for making clinical decisions by studying how to plot the knowledge graph for a specific disease using the available resources. Methods The top semantic concepts of disease,examination,operation,departments,drugs,symptoms,anatomy and the semantic relation between 2 types of attribute and 14 types of semantics were defined. The hierarchical relations were established and the synonymous terms were integrated using the available thesauri and ontology respectively according to their semantic concepts. The semantic relations were extracted and the synonymous terms were integrated using the resources in knowledge repositories. Results The knowledge graph for heart failure was plotted,which includes 524 concepts and 4360 semantic relations. Conclusion The knowledge graph for heart failure can be plotted using the available thesauri and ontology combined with knowledge integration technologies.