人工神经网络法用于高效毛细管电泳分离条件优化的研究

Separation Optimization in High Performance Capillary Electrophoresis by Artificial Neural Network

把人工神经网络（ＡＮＮ）法应用于高效毛细管电泳（ＨＰＣＥ）分离条件的优化，给出了反向传播（ＢＰ）的ＡＮＮ模型的具体算法。用正交试验法同时考察了缓冲溶液组成、浓度、ｐＨ值和有机添加剂浓度等实验因素对ＨＰＣＥ分离合成色素和防腐剂的影响，采用误差反向传播方法建立了有效的ＡＮＮ预测模型，预测最佳分离条件，获得了满意的分离结果。

Artificial neural network(ANN)was applied to the optimization of the separation conditions in HPCE and the program of back_propagation model was given.The effect of the buffer composition,concentration,pH value and ethanol concentration on the separation of food colors and antiseptics was examined by using orthogonal design.The prediction model of artificial neural network was set up and the optimum separation conditions were predicted successfully.