HKGB: An Inclusive, Extensible, Intelligent, Semi-auto-constructed Knowledge Graph Framework for Healthcare with Clinicians’ Expertise Incorporated
Yong Zhang,Ming Sheng,Rui Zhou,Ye Wang,Guangjie Han,Han Zhang,Chunxiao Xing,Jing Dong
DOI: https://doi.org/10.1016/j.ipm.2020.102324
2020-11-01
Abstract:<p>Health knowledge graph provides an ideal technical means to integrate heterogeneous data resources and enhance knowledge-based services. There are many challenges for the construction of health knowledge graph such as complex concepts and relationships, various medical standards, heterogeneous data structures, poor data quality, highly accurate and interpretable services, etc.</p><p>In this paper, firstly, we propose Health Knowledge Graph Builder (HKGB), an end-to-end platform which could be used to construct disease-specific and extensible health knowledge graphs from multiple sources. Secondly, we analyze the capabilities and requirements of clinicians, design the tasks to involve the clinicians and implement a clinician-in-the-loop toolset to integrate the clinicians prior knowledge into the construction of health knowledge graphs. Thirdly, we design an extensible mechanism to add new diseases to an existing knowledge graph. Fourthly, we present a quantitative effort estimation algorithm to quantitatively evaluate the effort of clinicians during the construction, and use it to calculate the workloads such as 44.27 person days for knee osteoarthritis domain. Finally, we have developed several knowledge graph based tools to facilitate real applications.</p>
computer science, information systems,information science & library science