基于可视化知识图谱的心电图特征分析

ECG Feature Analysis Based on visual knowledge graph

本文针对病态心电图种类繁多、变异极大和诊断错误率高的问题,构建了创新型的基于12导联病态心电图的封闭知识图谱,展示关于Neo4j和protege的可视化知识图谱成果。本文首先将心电图中的QRS波、P波等波形的数字属性值以及形态属性值抽取出来,进行语义类型及属性关系的分类。其次,将实体与关系导入知识图谱,利用OWL语言构建病系框架,形成完整的可视化知识图谱,使心电图诊断逻辑过程可视化。可推理的完善的封闭领域的医学知识图谱可视化应用系统是人工智能与医疗应用的完美结合,拥有巨大发展潜力,有望带来更廉价、高效、精准的医疗建议和诊断。

Aiming at the problems of various types、great variation and high diagnostic error rate of morbid ECG,this paper constructs an innovative closed knowledge graph based on 12 lead morbid ECG,and shows the results of visualization knowledge graph about Neo4j and protege.In this paper,we first extract the digital attribute value and morphological attribute value of QRS wave and P wave in ECG,and then classify the semantic types and attribute relations.Secondly,the entity and relationship are introduced into the knowledge graph,and OWL language is used to construct the disease system framework to form a complete visual knowledge graph,so as to visualize the logical process of ECG diagnosis.The reasonable and perfect visualization application system of medical knowledge graph in closed field is the perfect combination of artificial intelligence and medical application.It has great development potential and is expected to bring cheaper、efficient and accurate medical advice and diagnosis.