预测毛细管区带电泳有效淌度的支持向量回归建模方法

提出预测毛细管电泳迁移行为的支持向量回归建模方法。以核苷为实际研究对象，利用正交试验获得的数据，结合二标记物技术，用支持向量回归算法建立毛细管区带电泳的柱温、电压、缓冲液浓度和pH值与3种核苷的有效淌度之间的相关模型。将其与偏最小二乘回归和人工神经网络方法相比较，结果表明所建模型的预测准确性优于后两者，适宜用于毛细管电泳迁移行为的预测。

A Method for Building Calibration Model of Prediction of Electrophoretic Mobilities in Capillary Zone Electrophoresis Using Support Vector Regression

A novel method for building calibration model of prediction of electrophoretic mobilities in capillary zone electrophoresis using support vector regression (SVR) was proposed. Nucleotide was studied as an example and orthogonal design was applied to arrange the experiments. Support vector regression was used to build calibration models to predict the electrophoretic mobilities of nucleotide from separation conditions of capillary zone electorphresis, i.e. column temperature, voltage, buffer concentration and pH. The proposed method was compared with partial least square regression (PLSR) and artificial neural network (ANN) modeling methods. The results show that the predictive accuracy of calibration models built by SVR is better than that of the models built by PLSR and ANN. Therefore, the method is appropriate to predict electrophoretic mobilities in capillary zone electorphresis.