基于判别分析法对药房自动化发药机机内品种的优化

Optimization of varieties of drugs in automatic drug dispensing machine by discriminant analysis

目的优化药房自动化发药机机内品种,提高快速发药效率。方法以自动化发药机卡药品种的数字特征(药品包装的长度、宽度、高度、体积、重量)为分析因素,采用"步进法"对5个变量进行逐步判别分析,将有显著贡献的变量建立判别函数式,并使用回代估计法对函数式的准确率进行检验。结果筛选出影响卡药的主要因素为重量(weight)和高度(height),得到判别函数为F1=0.595 weight+0.588 height,F2=-0.932 weight+0.936 height,判别函数将测试样本数据分为3类(卡药、不卡药、不出药),其准确率为84.3%,验证样本的准确率为83%。结论采用判别分析法可以为药房自动化发药机机内品种的优化设计提供合理的统计学依据。

Objective To optimize varieties of drugs in automatic drug dispensing machine and improve efficiency of drug dispensing. Methods The numerical characteristics variables of fault drugs,which included length,width,height,weight and volume,were taken as analytical factors,the stepwise discriminant analysis was adopted to analyze those factors. The discrimination function was established with significant factors,and the accuracy of function formulas was tested with backward substitution method. Results The significant contributions screened out were weight and height,and the formulas were F1= 0.595 weight＋0.588 height,F2=-0.932 weight＋0.936 height. Discriminant function divided test sample data into 3 groups（stuck drugs,not stuck drugs,no drug）,its accuracy rate was 84.3%,the accuracy rate of verification sample was 83%. Conclusion Discriminant analysis can be used for the optimization of varieties of drugs in automatic drug dispensing machine.