Multivariate calibration of near-infrared spectra by using influential variables

摘要: Near-infrared (NIR) spectral analysis usually needs to take advantage of multivariate calibration. However, not all the variables in spectra have equal contributions a calibration model. Identification informative is key step build high performance According influence variable on model, influential (IV) defined and method for identification IVs proposed this work. In method, set partial least squares (PLS) models are built using subset selected randomly by Monte Carlo re-sampling, then clustering these investigated means principal component analysis. The that make grouping can be identified as IVs. Finally, PLS model with adopted Five NIR datasets used test applicability method. results show reasonable efficient enough produce accurate reliable predictions.