基于共词分析和可视化的高血压疾病关联性挖掘  

Hypertension related association mining based on co-word analysis and visualization

作　　者：刘莉[1] 姚京京 李俊 陈先来[3] 周宇葵[1] LIU Li;YAO Jingjing;LI Jun;CHEN Xianlai;ZHOU Yukui(School of Life Science,Central South University,Changsha 410013,China;Xiangya School of Stomatology,Central South University,Changsha 410008,China;Institute of Information Security and Big Data,Central South University,Changsha 410083,China)

机构地区：[1]中南大学生命科学学院,湖南长沙410013 [2]中南大学湘雅口腔医学院,湖南长沙410008 [3]中南大学信息安全与大数据研究院,湖南长沙410083

出　　处：《中国医学物理学杂志》2019年第5期614-620,共7页Chinese Journal of Medical Physics

基　　金：国家重点研发计划(2016YFC0901705)

摘　　要：目的:对高血压患者电子病历病案首页进行分析挖掘,揭示其中疾病诊断之间的关系。方法:以共词分析为基础,通过Python语言构建分析模块,采用Gephi复杂网络分析软件对结果进行展示。结果:基于3 632条电子病历记录,构建包含疾病诊断节点1 029个,共现关系边12 479条的疾病诊断共现网络,发现共现关系较强的疾病诊断集群。结论:从多角度、多层面对疾病诊断共现网络进行解读,并以可视化图谱的方式展示,揭示疾病诊断之间关系,为下一步构建更加完善的疾病图谱奠定基础。Objective To analyze and mine the electronic medical record home page of patients with hypertension, and to reveal the relationships among disease diagnoses. Methods Based on the co-word analysis, the analysis module was established by Python language, and the analysis results were displayed by Gephi complex network analysis software. Results Based on 3 362 electronic medical records, a disease diagnosis co-occurrence network containing 1 029 disease diagnosis nodes and 12 479 co-occurrence relationship lines was constructed, and a disease diagnosis cluster with strong co-occurrence relationship was found. Conclusion The disease diagnosis co-occurrence network can be interpreted from multi-angle and multi-level, and can be displayed in a visual map to reveal the relationships among disease diagnoses, laying a foundation for the further establishment of a more complete disease map.

关 键 词：电子病历 病案首页 高血压 共词分析 可视化 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] R312[自动化与计算机技术—计算机科学与技术]