摘要
[目的/意义]利用医学语义关系,基于医学文献开展主题演化类型与演化路径识别方法研究,为进一步丰富和完善主题演化理论与方法提供参考。[方法/过程]首先描述主题演化的新生、消亡、继承、分裂、融合和孤立6种类型,设计了基于统计与语义相结合的关键关联计算方法,提出利用密度和热度开展核心主题识别的研究,从关键关联与核心主题两方面共同识别主题的演化路径,进而确定主题演化类型。[结果/结论]采用本文研究思路与方法,对肝癌致病机制的研究领域文献开展主题演化路径识别与演化类型确定研究,实验证明所提方法可以较好描述领域的主要演化情况。后续研究将改进关键关联构建方法,以提升演化路径识别方法的精度。
[Purpose/significance] In order to enrich and improve topic evolution theories and methods,this paper studies the topic evolution type and method of path identification using medical semantic association based on medical literature.[Method/process] Firstly,the types of topic evolution are described,such as new born,disappearance,inheritance,fission,fusion and isolation.Then the key association calculation method based on the combination of statistics and semantics is designed,and the identification method of core topics using density and hotness is proposed.From the two aspects of key association and core topics,the paper can further identify the topic evolution path and type.[Result/conclusion] The proposed research method is applied in the research documents of liver cancer pathogenesis for topic evolution path identification and evolution type recognition.The experiment proves that the method can better describe the main evolution situation of the field.The follow-up research will improve the key association construction method to enhance the accuracy of the evoluation path recognition method.
出处
《情报理论与实践》
CSSCI
北大核心
2019年第3期114-119,共6页
Information Studies:Theory & Application
基金
国家自然科学基金项目“基于语义的医学领域前沿知识发现及演化机制研究”(项目编号:71303259)
国家自然科学基金项目“面向精准医学的基因-疾病-药物语义关系抽取研究”(项目编号:71704188)
中国医学科学院中央级科研院所基本科研业务专项(项目编号:2017PT63004)的研究成果之一。
关键词
主题演化
演化类型
演化路径
语义关联
topic evolution
evolution type
evolution path
semantic association
