医联网应用中异源健康数据语义融合方法研究

医疗物联网及移动医疗应用中多种传感器采集的生命体征数据，以及各类健康医疗数据彼此之间存在语义异构性，导致智能医疗物联设备数据融合困难。针对这一问题，研究了基于开放关联数据的语义消歧方法。首先对设备数据进行本体建模，形成局部本体；然后利用图匹配算法将局部本体与开放医疗关联数据进行概念对齐，间接消除异源数据间的语义异构性；最后，在运动手环与体重计数据融合实验中，通过与开放关联数据源的关联匹配判定血压和体重等异构概念属于语义相关概念。实验结果表明，通过与开放关联数据源关联，可以实现局部本体语义扩展，进一步实现异源医联网设备数据融合。

Heterogeneous healthcare data semantic fusion method for medical IoT application

Semantic heterogeneity exists in physiological index data collected by various sensors in medical Internet of Things and mobile medical applications, as well as in different kinds of healthcare/medical data. This leads to difficulty of data fusion for smart medical IoT devices. In order to solve this problem, this paper developed semantic disambiguation method based on open linked data. First, the method modeled device data to from local ontology. Then it used graph-matching algorithm to align concepts between local ontology and open medical data source to indirectly eliminate semantic heterogeneity of multi-source data. Finally, by linking and matching to open linked data sources, the experiment of data fusion between sports bracelets and intelligent weight meter equipment determine that heterogeneous concepts, such as blood pressure and weight, are semantic related concepts. The results show that linking to open data sources may realize semantic extension of local ontology, and then realize data fusion of heterogeneous medical IoT devices.