近红外光谱法用于纯化过程中鞣花酸含量的监控  

Application near-infrared spectroscopy in purification process of ellagic acid in monitoring

作　　者：李重阳[1] 武佳[1] 苏艳红[1] 白珂珂[1] 李晖[1] 

机构地区：[1]四川大学化学工程学院,四川成都610065

出　　处：《化学研究与应用》2013年第1期82-85,共4页Chemical Research and Application

摘　　要：采用近红外光谱技术结合遗传算法优化的小波神经网络,对大孔树脂纯化过程中橄榄果中的鞣花酸含量进行监控。通过小波变换对光谱进行去噪、压缩,作为人工神经网络的输入,同时以遗传算法优化神经网络的权值与阈值,并与常用的偏最小二乘(PLS)线性模型的建模效果进行比较。实验结果表明,两者都能够较准确的预测鞣花酸的含量,相对而言,人工神经网络(ANN)效果较好。The near-infrared spectroscopy combined with wavelet neural networks which is optimized by genetic algorithm was used to monitor the ellagic acid content in olive fruit in the macroporous resin purification process. After being denoised and compressed by wavelet transform,the spectrum was used as the input message of artificial neural network（ANN）. At the same time,the weights and threshold was optimized by genetic algorithm, and the results were compared with that of partial least squares （PLS）linear mod- el. Results of this experimental showed that both of the two methods couldpredict the content of ellagic acid, but artificial neural net- work is more accurately than partial least squares linear model

关 键 词：鞣花酸 近红外光谱 偏最小二乘 人工神经网络 

分 类 号：O657.3[理学—分析化学]