一种医学本体多层概念语义关联度度量模型研究  

Research on Semantic Association Metric of Multi-Level Concept Model for Medical Ontology

作　　者：刘玉文[1,2] 王凯[1] 徐济成[3] 

机构地区：[1]蚌埠医学院卫生管理系,安徽蚌埠233030 [2]中国科学技术大学计算机科学与技术学院 [3]安徽农业大学信息与计算机学院,安徽合肥230027

出　　处：《九江学院学报（自然科学版）》2017年第2期66-73,共8页Journal of Jiujiang University：Natural Science Edition

基　　金：安徽省高校自然科学一般项目(编号KJ2015B023by);蚌埠医学院自然科学重点项目(编号Byky1411ZD);安徽省教育厅人文社科重点项目(编号sk2015A405)的研究成果之一

摘　　要：数据语义关联度是大数据融合的核心环节。针对传统医学信息相似度计算模型的不足,文章提出一种基于语义关联度的多层医学信息概念语义距离度量模型,引入概念权重属性。通过领域专家对医学领域概念进行权重评分,综合考虑概念语义关系,通过加权概念实例闻、概念属性间以及概念与属性间的语义关系,并利用概念节点的最小近邻集搜索与其语义关系最接近的概念实体和结果概念分层思想,依次遍历概念层次树,获取概念相似集,最终计算出与目标概念语义距离最小的概念实体,达到概念融合的目的。该模型能够充分利用语义本体概念形式化表示以及分类层次关联度的优势,避免传统模型将概念集作为单一向量计算相似度的弊端,解决数据稀疏等问题。实验结果表明,多层概念语义关联度模型计算准确性较高。Data semantic correlation was a core part of data fusion. Other than using the traditional medical information similarity model deficiencies, we proposed a measurement model of multi-layer medical infor- mation coneept of semantic distance semantic correlation based on the introduction of concept of attribute weight. The weight score on the field of medicine concept was rate and the concept of semantic relations were considered, by experts in the field, Through the weighted concept instance, concept attribute and concept attributes and semantic relation, using the smallest neighbor concept node concept entity set search and its closest semantic relations, the target concept and semantie distance conceptual entity were calculated. This model could make full use of the seman- tic ontology formalization and classification level correlation advantages, avoided the traditional model of the concept set as single vector similarity calculation problems, and solved the problem of sparse data. The experimental results showed that the computational accuracy of the multi concept semantic relevance model was high.

关 键 词：医学信息度量 语义本体 语义关联 多层概念 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]