基于知识图谱的精神分裂症药物知识发现

采用路径发现、链路预测等方式从知识图谱中获取知识内在关联，挖掘知识间的隐含信息，为研究提供新方向。在利用多源数据库对获取的精神分裂症文献进行实体识别、关系抽取、属性补充的基础上，用Neo4j图数据库构建精神分裂症知识图谱，用Neo4j图数据库中的Cypher查询语句进行节点中心性挖掘和元路径查询，共发现67种药物，其中包含29种治疗精神分裂症药物和38种尚未用于治疗该疾病的药物。对未用于治疗精神分裂症的38种药物进行分析，发现这38种药物与精神分裂症存在隐含联系，可以从中挖掘治疗其的潜在药物。

Knowledge graph-based knowledge discovery of drugs for schizophrenia

The internal relationship of knowledge was detected from the knowledge graph with approach discovery method and link prediction method respectively, and the implicit information in knowledge was mined. The knowledge graph of schizophrenia was plotted by making use of the Neo4j graph database. The node centrality was mined and the meta-path was queried using the Cypher query language in Neo4j graph database based on the identified entities, extracted relationships between entities, and supplemented properties of entities in papers on schizoph-renia from the multisource database, which discovered 67 drugs, including 29 used in treatment of schizophrenia and 38 not used in treatment of schizophrenia. The 38 drugs not used in treat-ment of schizophrenia were analyzed, which showed an implicit relationship between these drugs and schizophrenia, from which certain potential drugs for the treatment of schizophrenia could be mined