雄伟
发表于 2025-3-23 09:42:50
Spectral Preprocessing Methods,In addition to the chemical information of the sample itself, the spectrum also contains some irrelevant information, such as electrical noise, sample background, and stray light.
thrombosis
发表于 2025-3-23 14:56:13
Wavelength Selection Methods,In the method of multivariate calibration in combination with spectra, the traditional view is that the multivariate calibration method (such as partial least squares, PLS) has strong anti-interference ability and can participate in the establishment of multivariate calibration model using the full spectrum.
tolerance
发表于 2025-3-23 19:26:22
Spectral Dimensionality Reduction Methods,The premise of multiple linear regression (MLR) is that the independent variables must be independent from each other. However, there is often a certain degree of correlation between the spectral variables, resulting in statistical multicollinearity.
Epithelium
发表于 2025-3-24 00:20:44
http://reply.papertrans.cn/23/2250/224920/224920_14.png
BOOST
发表于 2025-3-24 06:21:30
http://reply.papertrans.cn/23/2250/224920/224920_15.png
Cupidity
发表于 2025-3-24 08:44:14
Calibration Transfer Methods,A situation is often encountered in the process of spectral analysis that when the model established on one spectrometer (master, primary, and parent instrument) is used on another spectrometer (slave, child instrument), the model cannot give correct prediction results due to the difference in spectra measured by different instruments.
大都市
发表于 2025-3-24 13:31:49
http://reply.papertrans.cn/23/2250/224920/224920_17.png
现代
发表于 2025-3-24 17:49:20
Method of Selecting Calibration Samples,a robust model. As shown in Fig. ., the selection of samples is to select the row vector of spectral matrix . and the row vector of corresponding concentration matrix ., and the selection of wavelength is to select the column vector of spectral matrix ..
MORT
发表于 2025-3-24 22:26:53
http://reply.papertrans.cn/23/2250/224920/224920_19.png
CHIDE
发表于 2025-3-25 00:43:38
Multi-spectral Fusion Technology,er to obtain more comprehensive, more reliable, and richer characteristic data and then combine with chemometrics method to build regression or classification model for quantitative and qualitative analysis of samples.