基于主成分分析算法和K均值聚类算法的药品库存分类管理  

Drug Inventory Classification Management Based on PCA Algorithm and K-means Clustering Algorithm

作　　者：唐蕾[1] 邱磊 俞佳慧 冀召帅 TANG Lei;QIU Lei;YU Jiahui;JI Zhaoshuai(Department of Pharmacy,Beijing Tsinghua Changgeng Hospital,School of Clinical Medicine,Tsinghua University,Beijing 102218,China)

机构地区：[1]清华大学附属北京清华长庚医院/清华大学临床医学院药学部,北京102218

出　　处：《医药导报》2025年第4期682-686,共5页Herald of Medicine

摘　　要：目的针对目前药品分类主观性较强、分类标准模糊、影响因素繁杂的问题,探讨一种科学的药品分类方法,以降低库存成本,提高库存的有效性。方法在北京某三级医院2021—2022年历史数据中随机抽取700种药品为研究对象,通过主成分分析(PCA)算法和K均值聚类(K-means)算法对研究对象进行分类。结果确定轮廓系数为0.3470的分类数4为最佳分类数,将700种药品分为4类,其中有363种归为第一类,186种归为第二类,94种归为第三类,57种归为第四类。将该文研究的药品分类方法模拟运用到某三级医院2023年第二季度的药品库存管理中,模拟结果表明该分类方法能够降低库存成本,提高库存有效性。结论基于PCA算法和K-means聚类算法的药品分类方法能够为药品库存分类管理提供可靠依据。Objective To address the current issues of strong subjectivity in drug classification,vague classification standards,and complex influencing factors,this study aims to explore a scientific method for drug classification to reduce inventory costs and improve inventory effectiveness.Methods A total of 700 drugs were randomly selected from the historical data of a tertiary hospital in Beijing from 2021 to 2022 as the research subjects.The classification was conducted using the Principal Component Analysis(PCA)algorithm and the K-means clustering algorithm(K-means).Results The optimal number of classifications was determined to be 4,with a silhouette coefficient of 0.3470.The 700 drugs were divided into four categories,with 363 in the first category,186 in the second,94 in the third,and 57 in the fourth.The drug classification method studied in this paper was simulated and applied to the drug inventory management of a certain tertiary hospital in the second quarter of 2023.The simulation results indicated that the classification method studied in this paper could reduce inventory costs and improve inventory effectiveness.Conclusion The drug classification method based on PCA algorithm and K-means clustering algorithm can provide a reliable basis for the management of drug inventory classification.

关 键 词：药品分类 主成分分析算法 K均值聚类算法 药品库存管理 

分 类 号：R95[医药卫生—药学]