研究了应用人工神经网络进行粉末药品的无损定量分析。使用安体舒通粉末药品的近红外漫反射光谱数据建立人工神经网络模型,预测未知样品。讨论了影响网络的各参数,使用了逼近度作为网络新的评价标准。%The application of artificial neural network for nondestructive quantitative pharmaceutical analysis were investigated. A real data set from near -infrared diffuse reflectance spectra of spironolactone pharmaceutical powder was used to built up an artificial neural network model to predict unknown samples. The parameters affecting the network were discussed. A new network evaluation criterion, the degree of approximation, was employed.
展开▼