基于Logistic回归的短缺药品预警模型构建

Construction of Drug Shortage Warning Model Based on Logistic Regression

为了建立合理有效的短缺药品预警模型,本文使用K-S检验和Mann-Whitney U检验对指标进行筛选,对经过筛选处理后的预警指标进行因子分析,确保在消除指标间多重共线性的同时,保留尽可能多的初始指标信息。对提取的公因子使用Logistic回归分析,得到短缺药品预警模型,检验结果显示,训练样本和测试样本的预测正确率分别为97.06%和91.18%,表明该模型可以有效预测药品短缺。

In order to establish a reasonable and effective drug shortage warning model, K-S test and Mann-Whitney U test were used to filtrate the indicators, and factor analysis was performed on the early-warning indicators after filtrating, so as to ensure that the multicollinearity among indicators is eliminated and the maximum initial indicator information is retained as far as possible.Logistic regression was used for the extracted principal factors to obtain the warning model of drug shortage, and the test results show that the prediction accuracy of training samples and test samples are 97.06% and 91.18%,respectively, which indicates that the model can predict drug shortage effectively.