基于形式概念分析的面向药物适应症的领域本体构建方法  

An Approach to Building Drug Indication-oriented Domain Ontology Based on Formal Concept Analysis

作　　者：唐晓波[1,2] 伍星蓉[2] 

机构地区：[1]武汉大学信息系统研究中心,武汉430072 [2]武汉大学信息管理学院,武汉430072

出　　处：《情报杂志》2017年第12期159-165,172,共8页Journal of Intelligence

基　　金：国家自然科学基金项目"基于文本和Web语义分析的智能咨询服务研究"(编号:71673209)研究成果之一

摘　　要：[目的/意义]使用本体对医药信息进行组织和表示,不仅能够在语义的层面上有效地管理和利用医药知识,而且能够实现领域知识的融合共享,推动医药领域的智能应用。当前对药物本体的构建主要借助规范化的医学术语系统,忽略了药物适应症自然文本中隐含的药物语义关联。因此提出一种面向药物适应症的领域本体构建方法。[方法/过程]以"抗肿瘤药物"为例,对265条药物信息进行处理;将适应症文本进行语义补全,使用二次分词和合并的方法抽取特征词;利用形式概念分析,以"药物—适应症"的关联构建形式背景并生成概念格,根据概念格的结构对药物关系进行描述;使用OWL语言对概念格转化而来的本体进行描述。[结果/结论]对构建的本体进行一致性检验,并以可视化方式展示和分析。实验表明该构建方法能够表示药物与适应症的语义关联。[Purpose/Significance]Applying ontology to organize medical information can support managing and utilizing medical knowledge effectively in semantic meaning,which makes it possible to integrate and share various medical knowledge and promote the artificial application in medicine. The current methods of ontology building relied mainly on normalized medical terminology,however the implicit semantic relationships between drug indications in free text form were neglected. Based on such insufficiencies,we propose a drug indication-oriented method for domain ontology construction.[Methods/Process]We fetched 265 antineoplastics drugs through the web,and complemented indication texts according to its semantic meaning. After a twice-segmentation and merging method which was used to amend initial keywords,we built the concept context by the relationships between drugs and indications and then constructed a concept lattice. We applied OWL to represent our drug ontology according to the complicated concept lattice. Finally,the consistency and coherence of our ontology were tested,we also described this ontology in visual way.[Result/Conclusion] It is concluded that our method is available to indicate the semantic relationships between drugs and its indications.

关 键 词：形式概念分析 药物本体 药物适应症 本体构建 

分 类 号：G356[文化科学—情报学]