人工神经网络在近红外光谱建模中的应用及研究现状

Research and Application of ANN Method in Near Infrared Spectrum Modeling

对国内外近十年来人工神经网络在近红外光谱建模中的应用和研究进行了详细的综述,包括误差反向传播网络、径向基网络、支持向量机、自组织特征映射网、广义回归神经网络、概率神经网络、小波神经网络、模糊神经网络以及集成神经网络等的应用和研究。概括了这些网络的基本工作原理及优缺点。最后根据神经网络的发展方向和工农业的发展需求,提出了今后人工神经网络在近红外建模方面的发展方向。

The research and application of Artificial Neural Network （ANN） in the modeling of near in-frared spectrum at home and abroad in recent years are overviewed. The ANN includes back-propagation network, radiM-basis function network, support vector machine, self-organizing feature mapping, gener-alized regression neural network, probabilistic neural network, wavelet neural network, fuzzy network and neural network ensemble etc.. The basic operation principles, advantages and disadvantages of these networks are summarized. Finally, the trend of ANN in near infrared spectrum modeling in the future is proposed according to the development of ANN and the demands in industry and agriculture.