【作者】 张雅雄； 夏祥华； 柴宜民；

【Author】 Yaxiong Zhang~* Xianghua Xia Yimin Chai (School of Chemistry and Material Science,Shan’xi Normal University,Linfen, 041004,)

【机构】 山西师范大学化学与材料科学学院；

【摘要】 本研究工作采用基于遗传算法输入变量选择的径向基函数人工神经网络方法(RBFANN)用于重叠毛细管电泳峰中多组分的定量分析,结果表明RBFANN方法可有效解决重叠毛细管电泳峰中多组分定量分析的问题。通过采用本文建议的基于遗传算法的输入变量选择策略,对模拟数据及实验数据的定量分析结果的准确度与精密度均得以改善。通过对不同分离度的模拟数据以及完全重叠与部分重叠的实验数据的考察,所得结果仍支持上述结论。更多还原

【Abstract】 The application of radial basis function artificial neural networks(RBFANN) based on genetic input selection for quantification of the multi-component in unresolved peaks in capillary zone electrophoresis(CZE) is reported.An optimization strategy based on genetic algorithm(GA) for input variable selection of RBFANN was proposed.When the corresponding multi-component capillary electrophoresis(CE) peaks can not be resolved completely only by separation techniques,RBFANN based on genetic input selection was proved to be a suitable tool for quantification of the multiple components.Both the simulated and the experimental data were applied to evaluate the validity of the proposed method.The electrophoretograms recorded at selected wavelength were used as multivariate input data of RBFANN.Based on the successful evaluation of the proposed method applying simulated data,the study also shows that the applying of genetic input selection in RBFANN can improve the accuracy and precision of quantification in both completely and partially overlapped multi-component CE peaks to some extent.更多还原