溶剂系统的模式识别

Pattern Recognition of Solvent Systems

通过三层BP人工神经网络 ,对溶剂系统进行模式识别。以经过规格化处理的Xt、Xe、Xm、Xn、Xd这 5个溶剂性质的选择性参数作为判别指标 ,5类共 19种不同的溶剂作为训练样本 ,得到最佳网络参数。然后以十多种有代表性的溶剂作为待测样本进行了模式识别。研究结果表明 ,人工神经网络用于溶剂系统的模式识别 ,识别结果与实际一致。本文结果有助于色谱分析及分离技术中溶剂的选择。

By means of the back propagation training algorithm, three layer artificial neural network has been applied to pattern recognition of solvent systems.According to the five selective parameters of solvent Xt、Xe、Xm、Xn and Xd,which are disposed specifically, the 33 common solvents in which 19 solvents are training samples and others do test samples have been identified successfully. The results has been demonstrated that Pattern recognition of common solvent by artificial neural network agree with the fact.The Pattern recognition may be used to aid in the selection of solvents in chromatogram and other separation techniques.