呼吸系统疾病知识图谱的构建

Construction of Knowledge Graph of Respiratory System Disease

呼吸系统疾病是中国十大死亡率疾病之一。为解决其数据分布庞杂、难以深层次应用的问题,通过采用自顶向下的方法,构建了呼吸系统疾病知识图谱。首先,参考医学文献构建了呼吸疾病词典,并根据该疾病词典爬取了垂直医疗网站的呼吸疾病数据;然后,利用知识融合完成多源异构数据的链接,通过定义不同疾病属性相似度来计算疾病实体相似度,提高准确率;最后,将呼吸系统疾病数据存储到Neo4j中,并做可视化展示。

Respiratory diseases are one of the top ten mortality diseases in China.Therefore,in order to solve the problem that its data distribution is heterogeneous and difficult to apply in a deep level,the knowledge graph of respiratory diseases is construct-ed by adopting a top-down approach.First,a respiratory disease dictionary is constructed by referring to medical literature,includ-ing crawling the respiratory disease data from vertical medical websites based on this disease dictionary.Secondly,knowledge fusion is used to complete the linking of heterogeneous data from multiple sources,and the similarity of disease entities is calculated by de-fining the similarity of different disease attributes to improve the accuracy.In conclusion,the respiratory disease data are stored into Neo4j and visualized.