APPLICATION OF THE ANALYSIS METHOD COMBINING PRINCIPAL COMPONENT WITH NEURAL NETWORK IN NEAR-INFRARED SPECTRUM
Because of the serious overlapping of the near-infrared spectrum,it is necessary to use chemometrics technique to extract the related information.One of the core technologies of near-infrared spectrum analysis is the foundation about the function relation and the mathematical model between the spectrum information and the component.The principal component analysis and the artificial neural network are two chemometrics methods in common use,which has respective advantage and the limitation.Thus this paper put forward the arithmetic combination between neural network and principal component analysis in order to learn from their strong points to offset weakness, first,we compress original data into several independent variables by the principal component analysis, and then input the neural network which can reduce the network study time.on the basis of the research about basic principle and calculation step of the principal component analysis, the function in MATLAB are transferred in order to establish mathematical model; The compilation work of the training procedure is finished with the aid of the MATLAB neural network toolbox.
Principal Components Neural Network MA TLAB
Zhiyang Luo Hongxia Zhao Guanyun Xiao
Mechanical Engineering Department,Jinggangshan University Jian Jiangxi 343009,P.R.China
国际会议
2009 International Symposium on Computer Science and Technology(2009 中国宁波国际计算机科学与技术学术大会)
宁波
英文
80-84
2009-11-20（万方平台首次上网日期，不代表论文的发表时间）