基于主成分分析的BP神经网络在药品销售预测中的应用

Sale Volume Prediction of Drugs Using Back Propagation Neural Network Based on Principal Component Analysis

评价基于主成分分析的BP神经网络方法在药品销售预测中的可行性。采用主成分分析的方法对氯吡格雷、肝素钠、肝素钙等低分子肝素相关产品的销售额数据进行处理,形成新的指标体系,而后应用BP神经网络的方法建立模型,评价模型的拟合能力。结果:采用主成分分析的方法从各相关药物销售额数据中获得各主成分;使用BP神经网络建模并测试,测试结果误差较小。采用主成分分析的BP网络方法可以很好地对低分子肝素产品的销售额进行模拟。

To evaluate the use of PCA--BP ANN in the sale volume prediction of drugs, the method of principal component analysis was adopted to eliminate correlation of evaluation index of sample and a new system of evaluation index was formed on the basis of the initial one. Then it established a neural network mode to evaluate the use of PCA--BP ANN in the sale volume prediction of drugs. In the principal component analysis, the first three principal components account for about 80% of the total variability in the principal component analysis. And in the sale volume of low molecular-weight heparin, the Back Propagation Neural Network is effective. The results demonstrated that the back Propagation Neural Network based on principal component analysis can be well used in the sales volume prediction of low molecular-weight heparin in short time.