Combination of Modified Optical Path Length Estimation and Correction and Moving Window Partial Least Squares to Waveband Selection for the Fourier Transform Near-Infrared Determination of Pectin in Shaddock Peel

Improvement of NIR prediction ability by dual model optimization in fusion of NSIA and SA methods

Feature selection and sample partitioning are both important to establish a quantitative analytical model for near-infrared (NIR) spectroscopy. The classical interval partial least squares (iPLS) model for waveband selection can be improved in combination of the simulated annealing (SA) algorithm. The sample set partitioning based on a joint x-y distance (SPXY) method for sample partitioning is based on the distances of both the x- and y- dimensions; it is expected to be optimized using the non-dominant sorting strategies (NS) combined with the immune algorithm (IA). In this study, we investigated the dual model optimization mode for simultaneous selection of feature waveband and sample partitioning, and proposed a novel method defined as SA-iPLS & SPXY-NSIA. The method explores a population evolution process, and takes the candidate individual as the link for the fusion optimization of SA-iPLS and SPXY-NSIA. The method screens feature wavebands and observes a good partition of the modeling samples, to construct a combined optimization strategy for fusion optimization of the target waveband and suitable sets of sample partitioning. The performance of the SA-iPLS & SPXY-NSIA method was tested using a soil sample dataset. To prove model enhancement, the proposed method was compared to the two traditional methods of Kennard-Stone (KS) and SPXY in combination with SA-iPLS. Experimental results show that the fusion model established by SA-iPLS & SPXY-NSIA performed better than the KS-SA-iPLS and SPXY-SA-iPLS models. The best testing results of the fusion model is with RMSET, RPD_T and R_T observed as 0.0107, 1.7233 and 0.9097, respectively. The proposed method is prospectively able to effectively improve the predictive ability of the NIR analytical model.

Development of a Non-Destructive Method for Detection of the Juiciness of Pear via VIS/NIR Spectroscopy Combined with Chemometric Methods

Juiciness is a primary index of pear quality and freshness, which is also considered as important as sweetness for the consumers. Development of a non-destructive detection method for pear juiciness is meaningful for producers and sellers. In this study, visible-near-infrared (VIS/NIR) spectroscopy combined with different spectral preprocessing methods, including normalization (NOR), first derivative (FD), detrend (DET), standard normal variate (SNV), multiplicative scatter correction (MSC), probabilistic quotient normalization (PQN), modified optical path length estimation and correction (OPLECm), linear regression correction combined with spectral ratio (LRC-SR) and orthogonal spatial projection combined with spectral ratio (OPS-SR), was used for comparison in detection of pear juiciness. Partial least squares (PLS) regression was used to establish the calibration models between the preprocessing spectra (650-1100 nm) and juiciness measured by the texture analyzer. In addition, competitive adaptive reweighted sampling (CARS) was used to identify the characteristic wavelengths and simplify the PLS models. All obtained models were evaluated via Monte Carlo cross-validation (MCCV) and external validation. The PLS model established by 19 characteristic variables after LRC-SR preprocessing displayed the best prediction performance with external verification determination coefficient (R2v) of 0.93 and root mean square error (RMSEv) of 0.97%. The results demonstrate that VIS/NIR coupled with LRC-SR method can be a suitable strategy for the quick assessment of juiciness for pears.

Investigation of sample partitioning in quantitative near-infrared analysis of soil organic carbon based on parametric LS-SVR modeling

A framework for sample partitioning is proposed to take into account the tunable ratio of numbers of calibration and prediction samples, in consideration with the randomness, stability and robustness of calibration models.

FT-NIR spectroscopy and Whittaker smoother applied to joint analysis of duel-components for corn

Protein and total fat are two ingredients to measure the quality of corn. The aim of this study is to evaluate the quality of corn by the dual-component join determination through Fourier transform near infrared (FT-NIR) spectroscopic analysis. The calibration models were established by the systematic study performed respectively in the four regions of the whole range, the second overtone, the first overtone, and the combination. Whittaker smoother was introduced as an attractive alternative data preprocessing method. With the optimized parameters, Whittaker smoother indicates its priority for improving modeling results in any of the four regions. The predictive abilities were compared between the joint analysis of protein and total fat and the separate analysis of each single component by partial least squares (PLS) modeling. The uncertainty in parameter was further estimated for the linear models. It is suggested that the joint analysis of dual-component always leads to better predictive results, and also provided good evaluation results for the independent validation samples. For the joint analysis, the optimal region for protein was the combination (5400-4000 cm(-1)), and the optimal region for total fat was the first overtone (7200-5400 cm(-1)). The optimal PLS models also provided appreciate predictive performance for both protein and total fat. And the parameter uncertainty determination provided an acceptable estimate of the measured uncertainty for the FT-NIR analysis of corn. In general, the joint analysis of dual-component is a better strategy for FT-NIR analysis of corn, and it is hoped to be tested for other objects.