人工神经网络-近红外光谱法测定桉树中综纤维素的含量

采用近红外光谱技术结合广义回归神经网络(GRNN)建立测定桉树中综纤维素的定量分析模型。以72个桉树样品作为实验材料,对光谱数据进行平滑、求导、压缩以及归一化,用桉树的近红外光谱数据建立广义回归神经网络模型.预测模型的预测均方根误差为0.0198。结果表明,该方法测量比较准确,可以用于桉树中综纤维素含量的预测。

Determination of holocellulose content of eucalyptus by using artificial neural network and near-infrared spectroscopy

Using near-infrared reflectance spectroscopy combined with generalized regression neural network(GRNN),a model for determining holocellulose content of eucalyptus is established.72 eucalyptus samples are used as experimental material.The spectral data are pretreated by smoothing,derivative,compress and scaling.A real data set from nearinfrared reflectance spectroscopy of eucalyptus are used to build up models with GRNN. The root mean square error of predicted model are 0.0198.These results demonstrate that the method is precise.It can be used to determinate holocellulose content of eucalyptus.