摘要
文中将主成分分析和BP神经网络方法相结合,用于对近红外光谱数据进行预处理和回归分析,较好地解决了近红外分析中的非线性关联问题。实验结果表明,该方法在近红外光谱数据的分析中与传统的化学计量学方法相比有较好的应用效果。
Prineipal component analysis （PCA） and back- propagation neural network were used for pretreatrnent and analysis of the near infrared（NIR） spectroscopy data. The method solved the non - linearity relating problem in NIR analysis. Proved by example, the method used in this article can acquire better result than the traditional chemistry metrology methods in NIR analysis.
出处
《计算机技术与发展》
2006年第5期1-3,87,共4页
Computer Technology and Development
关键词
神经网络
近红外光谱分析
主成分分析法
neural network inear infrared spectroscopy analysis
principal component analysis
