基于条件概率的临床诊疗事件打包算法研究

发布时间：2018-03-05 20:26

  本文选题：过程挖掘　切入点：临床诊疗过程　出处：《计算机集成制造系统》2017年05期 　论文类型：期刊论文

【摘要】：为了从过往的医疗数据中得到清晰可理解的过程模型,并将其合理应用于医疗决策的制定以及临床路径的改善,提出一种全新的临床诊疗过程挖掘方案,通过将相关临床诊疗事件进行合理打包来减少作为过程挖掘算法输入的事件个数,从而简化挖掘到的临床诊疗过程模型。针对临床诊疗事件打包,提出一种基于条件概率的打包算法,该算法将条件概率作为衡量事件之间关联程度的标准,并将关联程度达到一定程度的事件进行打包。实验结果表明,所提出的临床诊疗过程挖掘方案确实能够得到清晰可理解的过程模型,所提出的打包算法能够在更高容忍度的基础上得到更加精确、合理的结果。
[Abstract]:In order to get a clear and understandable process model from the past medical data and apply it to the making of medical decision making and the improvement of clinical path, a new mining scheme of clinical diagnosis and treatment process is proposed. By reasonably packaging the relevant clinical diagnosis and treatment events to reduce the number of events as input of the process mining algorithm, the model of clinical diagnosis and treatment process can be simplified. A packing algorithm based on conditional probability is proposed, in which conditional probability is taken as the criterion to measure the degree of correlation between events, and events with a certain degree of association are packaged. The experimental results show that, The proposed mining scheme of clinical diagnosis and treatment process can indeed obtain a clear and understandable process model, and the proposed packaging algorithm can obtain more accurate and reasonable results on the basis of higher tolerance.
【作者单位】： 清华大学软件学院;
【分类号】：R197.3;TP311.13
，

本文编号：1571749