【摘要】 报道了一种基于卡尔曼滤波的线性神经网络用于 B族维生素混合组分的定量同时测定 ,用相关系数和标准偏差从原始紫外光谱数据中挑选 1 1个波长点供线性神经网络处理 ,以极少的迭代次数达到较高的精度 ,预测结果较好 ,相关系数大于 0 .99983 ,标准偏差 0 .1 1 3 0 1更多还原

【Abstract】 Simultaneous determination of the multicomponents of four B-group vitamins with a novel Linear Neural Network (LNN) algorithm was reported. Based on the data correlation coefficient and standard deviation, 11 representative wavelength points were selected from the original UV spectral data (240 points) as the original input patterns for FLN to build an NN model. The results obtained by LNN is satisfactory, with a correlation coefficient 0.999 83 and a standard deviation 0.113 01.更多还原