作者: U. Depczynski , K. Jetter , K. Molt , A. Niemöller
DOI: 10.1016/S0169-7439(98)00208-1
关键词: Fourier series 、 Chemometrics 、 Biological system 、 Wavelet transform 、 Regression 、 Wavelet 、 Principal component regression 、 Fast Fourier transform 、 Statistics 、 Mathematics 、 Calibration (statistics)
摘要: Abstract In this paper, we present wavelet coefficient regression (WCR) in combination with a genetic algorithm (GA) as method for multicomponent analysis by Near Infrared Spectrometry. The results are compared other multivariate calibration methods like Fourier (FCR), principal component (PCR) and absorbance value at selected wavelengths (AVR). It is shown that comparison to conventional methods, WCR quite unique the fact it self-adaptive. This means steps of pretreatment, selection specific wavelength regions performed automatically one step.