基于模糊聚类的医保报销数据异常检测方法  

Anomaly Detection Method for Medical Insurance Reimbursement Data Based On Fuzzy Clustering

作　　者：廖黎 尚思琪 聂娟娟 LIAO Li;SHANG Si-qi;NIE Juan-juan(Wuhan Hospital of Integrated Traditional Chinese andWestern Medicine(Wuhan No.1 Hospital),Wuhan 430022 China)

机构地区：[1]武汉市中西医结合医院(武汉市第一医院),湖北武汉430022

出　　处：《自动化技术与应用》2024年第4期151-154,共4页Techniques of Automation and Applications

摘　　要：在检测医保报销数据异常工作中,常见的检测方法在处理非线性数据时,数据分类精度不足,导致检测精度不足。针对这一问题,提出基于模糊聚类的医保报销数据异常检测方法。采用数据预处理技术和统计技术,将各类医保报销数据汇总,经过计算得到各数据集的相关矩阵。使用模糊聚类技术对数据特征模糊处理,得到各个数据集的模糊熵,以模糊熵作为条件,筛选出合适的数据特征。结合数据集的相关特征,定义数据检测条件,完成数据异常检测。实验结果表明:所提检测方法在处理非线性数据上分类精度高,检测评估中调和水平高于常见的检测方法。In the detection of abnormal medical insurance reimbursement data,the common detection methods have insufficient data classification accuracy when dealing with nonlinear data,resulting in insufficient detection accuracy.To solve this problem,an anomaly detection method of medical insurance reimbursement data based on fuzzy clustering is proposed.Using data preprocessing technology and statistical technology,it summarizes all kinds of medical insurance reimbursement data,and calculates the correlation matrix of each data set.Fuzzy clustering technology is used to deal with the data features,and the fuzzy entropy of each data set is obtained.Taking the fuzzy entropy as the condition,the appropriate data features are selected.Combined with the relevant characteristics of the data set,the data detection conditions are defined to complete the data anomaly detection.The experimental results show that the classification accuracy of the detection method is high in dealing with nonlinear data,and the harmonic level in detection and evaluation is higher than that in common detection methods.

关 键 词：模糊聚类 医保系统 报销数据 数据检测 数据特征 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]