基于改进 k -means的医院患者流向异常聚类研究

Research on Cluster of Abnormal Flow Direction of Hospital Patients Based on Improved k-means

针对医院患者流向异常监测需求,提出一种改进k-means的患者流向异常监测模型。首先,采用两阶段k-means对传统的k-means聚类进行改进,以解决初始质心随机选择问题;其次,构建改进k-means的医院患者流向异常检测模型,并通过仿对改进k-means算法聚类和患者流向异常检测进行验证。结果表明,这种基于改进k-means算法的医院患者流向异常检测模型,可有效实现流向患者聚类,且聚类时间和聚类效果优于传统的CobWeb、DBScan、Hierarchical Cluster聚类算法,有效提高了医院患者流量检测低的问题。

Aiming at the needs of monitoring abnormal flow of patients in hospital,an improved k-means model for monitoring abnormal flow of patients is proposed.Firstly,two-stage k-means is used to improve the traditional k-means clustering to solve the problem of random selection of initial centroid.Secondly,an improved k-means model for detecting abnormal flow direction of hospital patients is constructed,and the improved k-means algorithm clustering and abnormal flow direction detection of patients are verified by simulation.The results show that the hospital patients flow anomaly detection model based on the improved k-means algorithm can effectively achieve the flow of patients clustering,and the clustering time and clustering effect are better than the traditional CobWeb,DBScan,hierarchical cluster algorithm,which can effectively improve the problem of low detection of hospital patients flow.