Partial Least Squares Regression Technique Research Articles

Prediction of octane numbers for commercial gasoline using distillation curves: a comparative regression analysis between principal component and partial least squares methods

In this study, it was found that limiting the number of explanatory distillation curves input variables in a multivariate regression method to the most significant ones (10%, 50%, 90%, and 100% of recovered condensate volumes) provides a reliable model that can be used as a quick and inexpensive method for determining the octane rating of commercial gasoline. Octane numbers were predicted using a combination of distillation curves and multivariate calibration techniques of partial least squares regression (PLSR) and principal component regression (PCR). The root mean square error of prediction (RMSEP) evaluation for calibration and prediction sets for PLSR technique varies between 0.545 and 0.902 for MON and RON, respectively, which is relatively acceptable when compared to earlier studies. Comparing PLSR and PCR results revealed that PCR has slightly lower RMSEP values for MON and RON, 0.527 and 0.899, respectively. The fact that measurements of studied key temperatures are a routine part of gasoline quality control makes the current technique a promising alternative to the suggested PLS distillation curves models in the literature, which require a larger number of explanatory variables. This method can also replace the costly and time-consuming spectral analysis now used to determine the octane rating of gasoline.

Soil Chemical Properties and Fire Severity Assessment Using VNIR Proximal Spectroscopy in Fire-Affected Abandoned Orchard of Mediterranean Croatia

This paper aims to evaluate the ability of VNIR proximal soil spectroscopy to determine post-fire soil chemical properties and discriminate fire severity based on soil spectra. A total of 120 topsoil samples (0–3 cm) were taken from 6 ha of unburned (control (CON)) and burned areas (moderate fire severity (MS) and high fire severity (HS)) in Mediterranean Croatia within one year after the wildfire. Partial least squares regression (PLSR) and an artificial neural network (ANN) were used to build calibration models of soil pH, electrical conductivity (EC), CaCO3, plant-available phosphorus (P2O5) and potassium (K2O), soil organic carbon (SOC), exchangeable calcium (exCa), magnesium (exMg), potassium (exK), sodium (exNa), and cation exchange capacity (CEC), based on soil reflectance data. In terms of fire severity, CON samples exhibited higher average reflectance than MS and HS samples due to their lower SOC content. The PCA results pointed to the significance of the NIR part of the spectrum for extracting the variance in reflectance data and differentiation between the CON and burned area (MS and HS). DA generated 74.2% correctly classified soil spectral samples according to the fire severity. Both PLSR and ANN calibration techniques showed sensitivity to extract information from soil features based on hyperspectral reflectance, most successfully for the prediction of SOC, P2O5, exCa, exK, and CEC. This study confirms the usefulness of soil spectroscopy for fast screening and a better understanding of soil chemical properties in post-fire periods.

Improving the accuracy of soil organic carbon content prediction based on visible and near-infrared spectroscopy and machine learning

Choosing appropriate multivariate calibration and preprocessing transformation techniques is important in the determination of soil organic carbon (SOC) content based on visible and near-infrared (Vis–NIR) spectroscopy. The performance levels of partial least-squares regression (PLSR), support vector machine regression (SVMR), and wavelet neural network (WNN) calibration methods coupled with different preprocessing approaches were compared using three kinds of criteria, including the coefficient of determination (R2), root mean square error (RMSE), and residual prediction deviation (RPD). A total of 328 soil samples collected from the south bank of Hangzhou Bay were used as the dataset for the calibration–validation procedure and SOC content inversion. The effects of spectra preprocessing transformation methods were evaluated for raw spectra, Savitzky–Golay smoothing with the first derivatives of reflectance (FDR) and Savitzky–Golay smoothing with logarithm of reciprocal of the reflectance (log R−1). The results indicate that the SVMR is superior to the PLSR, and WNN models for SOC content prediction. The combination of the SVMR model with FDR provided the best prediction results for the SOC content, with R2p = 0.92, RPDP = 2.82, RMSEP = 0.36%, and a kappa correlation coefficient of interpolation as high as 0.97. The FDR of Vis–NIR spectroscopy combined with the SVMR model is recommended over the PLSR and WNN modeling techniques for the high-accuracy determination of the SOC content.

Estimation of the Yield and Plant Height of Winter Wheat Using UAV-Based Hyperspectral Images.

Crop yield is related to national food security and economic performance, and it is therefore important to estimate this parameter quickly and accurately. In this work, we estimate the yield of winter wheat using the spectral indices (SIs), ground-measured plant height (H), and the plant height extracted from UAV-based hyperspectral images (HCSM) using three regression techniques, namely partial least squares regression (PLSR), an artificial neural network (ANN), and Random Forest (RF). The SIs, H, and HCSM were used as input values, and then the PLSR, ANN, and RF were trained using regression techniques. The three different regression techniques were used for modeling and verification to test the stability of the yield estimation. The results showed that: (1) HCSM is strongly correlated with H (R2 = 0.97); (2) of the regression techniques, the best yield prediction was obtained using PLSR, followed closely by ANN, while RF had the worst prediction performance; and (3) the best prediction results were obtained using PLSR and training using a combination of the SIs and HCSM as inputs (R2 = 0.77, RMSE = 648.90 kg/ha, NRMSE = 10.63%). Therefore, it can be concluded that PLSR allows the accurate estimation of crop yield from hyperspectral remote sensing data, and the combination of the SIs and HCSM allows the most accurate yield estimation. The results of this study indicate that the crop plant height extracted from UAV-based hyperspectral measurements can improve yield estimation, and that the comparative analysis of PLSR, ANN, and RF regression techniques can provide a reference for agricultural management.

Objective measures of grape quality: From Cabernet Sauvignon grape composition to wine sensory characteristics

In an investigation of objective measures that link grape composition to wine quality, this study sought to identify Cabernet Sauvignon grape parameters that predict the sensory properties of the corresponding wines. Eleven chemical measures comprising volatile and non-volatile compounds, enzyme activity plus standard industry harvest measurements were applied to grape samples obtained from different regions throughout South Eastern Australia over three vintages. Grapes underwent controlled vinification and the resulting wines evaluated with sensory descriptive analysis. The entire multi-vintage data sets were combined and modelled using a combination of partial least squares (PLS) and sequential and orthogonalised (SO) -PLS regression techniques. Optimal models were obtained with single sensory attributes rather than global modelling with the entire sensory profile. Five grape chemical measures, which in the main were harvest parameters, were used along with colour, total phenolics and tannin, targeted volatiles, and flavonols, and orthogonalised to model 14 sensory attributes of the Cabernet Sauvignon wines. The seven remaining measures were not used due to their poor ability to model wine sensory attributes, with enzyme activity and tannin by HPLC explaining the least. The study revealed new insights into the relationship between grape chemistry and wine sensory characters, which has implications for developing an objective measurement system for determining grape quality.

Determination of creatinine level in patient blood samples by Fourier NIR spectroscopy and multivariate analysis in comparison with biochemical assay

In this paper, in vivo spectra from 23 patients’ blood samples with various Creatinine (Cr) concentration levels ranging from 0.96 to 12.5 mg/dL were measured using Fourier transform near-infrared spectrometer (FT-NIRS) and spectrum quantitative analysis method. Since Cr undergoes passive filtration, it serves as a key biomarker of kidneys function via the estimation of glomerular filtration rate. Thus, increased blood Cr concentration reflects impaired renal function. After spectra pre-processing and outlier exclusion, a spectral model was developed based on partial least squares regression (PLSR) method, wherein Cr concentrations correlated with filtered NIR spectra across several peaks, where Cr is known to absorb NIR light. Several statistical metrics were applied to estimate the model efficiency during data analysis. Comparison of spectra-derived concentrations to reference Cr measurements by the current gold-standard Jaffe’s method held in hospital lab revealed a Cr prediction accuracy of 1.64[Formula: see text]mg/dL with good correlation of [Formula: see text]. Bland-Altman plots were used to compare between our calculations and reference lab values and reveal minimal bias between the two. The finding presented the potential of FT-NIRS coupled with PLSR technique for Cr determination.

Estimating Chlorophyll Content of Leafy Green Vegetables from Adaxial and Abaxial Reflectance.

As a primary pigment of leafy green vegetables, chlorophyll plays a major role in indicating vegetable growth status. The application of hyperspectral remote sensing reflectance offers a quick and nondestructive method to estimate the chlorophyll content of vegetables. Reflectance of adaxial and abaxial leaf surfaces from three common leafy green vegetables: Pakchoi var. Shanghai Qing (Brassica chinensis L. var. Shanghai Qing), Chinese white cabbage (Brassica campestris L. ssp. Chinensis Makino var. communis Tsen et Lee), and Romaine lettuce (Lactuca sativa var longifoliaf. Lam) were measured to estimate the leaf chlorophyll content. Modeling based on spectral indices and the partial least squares regression (PLS) was tested using the reflectance data from the two surfaces (adaxial and abaxial) of leaves in the datasets of each individual vegetable and the three vegetables combined. The PLS regression model showed the highest accuracy in estimating leaf chlorophyll content of pakchoi var. Shanghai Qing (R2 = 0.809, RMSE = 62.44 mg m−2), Chinese white cabbage (R2 = 0.891, RMSE = 45.18 mg m−2) and Romaine lettuce (R2 = 0.834, RMSE = 38.58 mg m−2) individually as well as of the three vegetables combined (R2 = 0.811, RMSE = 55.59 mg m−2). The good predictability of the PLS regression model is considered to be due to the contribution of more spectral bands applied in it than that in the spectral indices. In addition, both the uninformative variable elimination PLS (UVE-PLS) technique and the best performed spectral index: MDATT, showed that the red-edge region (680–750 nm) was effective in estimating the chlorophyll content of vegetables with reflectance from two leaf surfaces. The combination of the PLS regression model and the red-edge region are insensitive to the difference between the adaxial and abaxial leaf structure and can be used for estimating the chlorophyll content of leafy green vegetables accurately.

Evaluation of Grass Quality under Different Soil Management Scenarios Using Remote Sensing Techniques

Hyperspectral and multispectral imagery have been demonstrated to have a considerable potential for near real-time monitoring and mapping of grass quality indicators. The objective of this study was to evaluate the efficiency of remote sensing techniques for quantification of aboveground grass biomass (BM) and crude protein (CP) in a temperate European climate such as Ireland. The experiment was conducted on 64 plots and 53 paddocks with varying quantities of nitrogen applied. Hyperspectral imagery (HSI) and multispectral imagery (MSI) were analyzed to develop the prediction models. The MSI data used in this study were captured using an unmanned aircraft vehicle (UAV) and the satellite Sentinel-2, while the HSI data were obtained using a handheld hyperspectral camera. The prediction models were developed using partial least squares regression (PLSR) and stepwise multi-linear regression (MLR). Eventually, the spatial distribution of grass biomass over plots and paddocks was mapped to assess the within-field variability of grass quality metrics. An excellent accuracy was achieved for the prediction of BM and CP using HSI (RPD > 2.5 and R2 > 0.8), and a good accuracy was obtained via MSI-UAV (2 < RPD < 2.5 and R2 > 0.7) for the grass quality indicators. The accuracy of the models calculated using MSI-Sentinel-2 was reasonable for BM prediction and insufficient for CP estimation. The red-edge range of the wavelengths showed the maximum impact on the predictability of grass BM, and the NIR range had the greatest influence on the estimation of grass CP. Both the PLSR and MLR techniques were found to be sufficiently robust for spectral modelling of aboveground BM and CP. The PLSR yielded a slightly better model than MLR. This study suggested that remote sensing techniques can be used as a rapid and reliable approach for near real-time quantitative assessment of fresh grass quality under a temperate European climate.

High-throughput field phenotyping using hyperspectral reflectance and partial least squares regression (PLSR) reveals genetic modifications to photosynthetic capacity

Spectroscopy is becoming an increasingly powerful tool to alleviate the challenges of traditional measurements of key plant traits at the leaf, canopy, and ecosystem scales. Spectroscopic methods often rely on statistical approaches to reduce data redundancy and enhance useful prediction of physiological traits. Given the mechanistic uncertainty of spectroscopic techniques, genetic modification of plant biochemical pathways may affect reflectance spectra causing predictive models to lose power. The objectives of this research were to assess over two separate years, whether a predictive model can represent natural and imposed variation in leaf photosynthetic potential for different crop cultivars and genetically modified plants, to assess the interannual capabilities of a partial least square regression (PLSR) model, and to determine whether leaf N is a dominant driver of photosynthesis in PLSR models. In 2016, a PLSR analysis of reflectance spectra coupled with gas exchange data was used to build predictive models for photosynthetic parameters including maximum carboxylation rate of Rubisco (Vc,max), maximum electron transport rate (Jmax) and percentage leaf nitrogen ([N]). The model was developed for wild type and genetically modified plants that represent a wide range of photosynthetic capacities. Results show that hyperspectral reflectance accurately predicted Vc,max, Jmax and [N] for all plants measured in 2016. Applying these PLSR models to plants grown in 2017 resulted in a strong predictive ability relative to gas exchange measurements for Vc,max, but not for Jmax, and not for genotypes unique to 2017. Building a new model including data collected in 2017 resulted in more robust predictions, with R2 increases of 17% for Vc,max. and 13% Jmax. Plants generally have a positive correlation between leaf nitrogen and photosynthesis, however, tobacco with reduced Rubisco (SSuD) had significantly higher [N] despite much lower Vc,max. The PLSR model was able to accurately predict both lower Vc,max and higher leaf [N] for this genotype suggesting that the spectral based estimates of Vc,max and leaf nitrogen [N] are independent. These results suggest that the PLSR model can be applied across years, but only to genotypes used to build the model and that the actual mechanism measured with the PLSR technique is not directly related to leaf [N]. The success of the leaf-scale analysis suggests that similar approaches may be successful at the canopy and ecosystem scales but to use these methods across years and between genotypes at any scale, application of accurately populated physical based models based on radiative transfer principles may be required.

Sediment yield is closely related to lithology and landscape properties in heterogeneous karst watersheds

As one of the largest contiguous karst regions in the world, Southwest China is vulnerable to soil erosion because of its unique geology and inappropriate land use. Understanding the linkages between geological conditions, land cover patterns and sediment yield is thus of paramount importance to effectively control sediment delivery in heterogeneous karst watersheds. The objective of this study was to examine how the specific sediment yield was related to lithology and landscape variables for 40 karst watersheds in southwest China. Due to these watershed variables were highly codependent, partial least squares regression (PLSR) was applied to elucidate the relationship between the sediment yield and 30 lithology and landscape variables. The first-order controls were determined by calculating the variable importance for the projection (VIP). Predictors with VIP scores larger than 1 were considered to be the most influential predictors of the PLSR model. Results indicated that the lithology and landscape indeed exerted substantial influence on sediment yield (explained variance: 66.8%) for the investigated 40 watersheds. Specifically, the first dominant factor for sediment yield was karst coverage (VIP = 2.19), followed by dolomite coverage (VIP = 1.42), landscape shape index (VIP = 1.33), patch richness (VIP = 1.23), relative patch richness (VIP = 1.23), number of patches (VIP = 1.22), and Simpson’s diversity index (VIP = 1.03). The PLSR technique enabled the elimination of codependency among the variables and facilitated a more unbiased view of the control factors for sediment yield. Therefore, this approach is recommended as a useful tool for identifying the dominant factors controlling sediment yield in heterogeneous karst watersheds.

Developing an intelligent system for the prediction of soil properties with a portable mid-infrared instrument

Portable mid-infrared (MIR) technology is well suited for the provision of detailed and inexpensive information on key soil properties for optimum soil management. This technology requires prior complex multivariate modelling. In this manuscript, we propose an intelligent system approach based on portable MIR spectroscopy data modelled by machine learning techniques to predict total carbon (TC), nitrogen (TN), cation exchange capacity (CEC), clay, silt and exchangeable sodium (Na+) in 458 representative soil samples from Australia. To find the best performing algorithm to be incorporated into the intelligent system, we evaluated the performance of a number of machine learning techniques, including Gaussian process regression (GPR), Random Forest (RF), M5 rules, and Bagging and Decision Trees, which were compared with traditional Partial Least Squares Regression (PLSR). Generally, all the models showed a satisfactory performance. However, GPR stood out (according to statistical tests), especially for the predictions of clay, Na+ and CEC, obtaining root mean squared error differences between the GPR and PLSR techniques of −0.27%, −0.40% and −0.63 cmol+ kg−1, respectively. This study evidenced the feasibility of using the GPR technique to model data provided by a portable MIR instrument for the prediction of soil properties, providing the best, most stable and reliable results.

Modeling of Water Holding Capacity Using Readily Available Soil Characteristics

Soil water holding capacity (WHC) and its spatial variability is heavily affected by soil organic matter and texture and had significant influence for varied application such as regulating plant growth, soil drainage and soil functional attributes. The present study was conducted in central region of the state of Arunachal Pradesh with the aim of modeling the WHC using readily available soil characteristics. Soil parameters were analyzed using standard methodologies, and WHC modeling was done by developed predictive model using partial least squares regression (PLSR) technique. Water holding capacity (WHC) and bulk density decrease with the increase in altitude. Soil was acidic in nature, and acidity decreases with increasing altitude. Soil texture ranges from sandy loam to sandy clay loam. There was a significant increase in porosity with the increase in altitude and decrease in soil organic matter. The developed models using PLSR technique showed good predictivity based on different statistical performances and error indices ranged R2 = 0.73–0.77; root-mean-square error (RMSE) = 5.69–6.08%; and mean squared error (MSE) = 32.45–37.02%). Variable importance in prediction analysis reflects the relative importance of each soil variables in the developed prediction models. It showed that clay was among the most important influencing variable of WHC followed by soil moisture, altitude and silt. Hence, clay percent can be considered among the most important variable in the model for WHC prediction. Correlation analysis of the variables also showed that water holding capacity was found to be strongly positively correlated (r2 = 0.88) with clay content followed by bulk density (r2 = 0.62) and organic matter (r2 = 0.43). However, a negative relationship was observed with other soil parameters. Based on the analysis of the results, the PLSR developed model results good fit with the selected variables. Hence, it may be concluded that the developed model could be successfully used to predict WHC using identified predictors under limiting data conditions. The findings showed altitude as one of the suitable predictors along with RASCs, as inclusion of this resulted in improvement in predictive accuracy of developed models of WHC. Hence, inclusion of altitude as predictor is being recommended for further studies.

A Comparison of Regression Techniques for Estimation of Above-Ground Winter Wheat Biomass Using Near-Surface Spectroscopy

Above-ground biomass (AGB) provides a vital link between solar energy consumption and yield, so its correct estimation is crucial to accurately monitor crop growth and predict yield. In this work, we estimate AGB by using 54 vegetation indexes (e.g., Normalized Difference Vegetation Index, Soil-Adjusted Vegetation Index) and eight statistical regression techniques: artificial neural network (ANN), multivariable linear regression (MLR), decision-tree regression (DT), boosted binary regression tree (BBRT), partial least squares regression (PLSR), random forest regression (RF), support vector machine regression (SVM), and principal component regression (PCR), which are used to analyze hyperspectral data acquired by using a field spectrophotometer. The vegetation indexes (VIs) determined from the spectra were first used to train regression techniques for modeling and validation to select the best VI input, and then summed with white Gaussian noise to study how remote sensing errors affect the regression techniques. Next, the VIs were divided into groups of different sizes by using various sampling methods for modeling and validation to test the stability of the techniques. Finally, the AGB was estimated by using a leave-one-out cross validation with these powerful techniques. The results of the study demonstrate that, of the eight techniques investigated, PLSR and MLR perform best in terms of stability and are most suitable when high-accuracy and stable estimates are required from relatively few samples. In addition, RF is extremely robust against noise and is best suited to deal with repeated observations involving remote-sensing data (i.e., data affected by atmosphere, clouds, observation times, and/or sensor noise). Finally, the leave-one-out cross-validation method indicates that PLSR provides the highest accuracy (R2 = 0.89, RMSE = 1.20 t/ha, MAE = 0.90 t/ha, NRMSE = 0.07, CV (RMSE) = 0.18); thus, PLSR is best suited for works requiring high-accuracy estimation models. The results indicate that all these techniques provide impressive accuracy. The comparison and analysis provided herein thus reveals the advantages and disadvantages of the ANN, MLR, DT, BBRT, PLSR, RF, SVM, and PCR techniques and can help researchers to build efficient AGB-estimation models.

Partial least squares regression (PLSR) associated with spectral response to predict soil attributes in transitional lithologies

ABSTRACTNew techniques and improvements are required to quantify soil’s chemical and physical properties on production environment, reducing environmental impacts and minimizing soil analysis time. The aim of this study is to evaluate the possibility to estimate the content of silt, sand, clay, total iron and organic matter in soils formed by different lithologies in Parana State, Brazil, using VIS-NIR spectrum associated with Partial Least Square Regression (PLSR). 200 soil samples were collected in an area formed by Lixisols, Cambisols, Ferralsols, Arenosols and Nitisols in a depths of 0–0.2 and 0.2–0.8 m. Spectral readings were obtained in laboratory by FieldSpec 3 JR sensor. The spectral curves of the samples were correlated to the attributes through PLSR. The results obtained for sand in prediction were better when compared to the other attributes, presenting R2 = 0.90, r = 0.95 and RPD = 2.3. Clay and total iron presented satisfactory results, mainly for RPD values, which were above 2.4. Based on the results, it can be concluded that the PLSR technique associated with the spectral response of the soils, was able to estimate sand, clay and total iron with accuracy in a region formed by reworked materials, derived from several lithologies.

Optimal Regression Method for Near-Infrared Spectroscopic Evaluation of Articular Cartilage.

Near-infrared (NIR) spectroscopy has been successful in nondestructive assessment of biological tissue properties, such as stiffness of articular cartilage, and is proposed to be used in clinical arthroscopies. Near-infrared spectroscopic data include absorbance values from a broad wavelength region resulting in a large number of contributing factors. This broad spectrum includes information from potentially noisy variables, which may contribute to errors during regression analysis. We hypothesized that partial least squares regression (PLSR) is an optimal multivariate regression technique and requires application of variable selection methods to further improve the performance of NIR spectroscopy-based prediction of cartilage tissue properties, including instantaneous, equilibrium, and dynamic moduli and cartilage thickness. To test this hypothesis, we conducted for the first time a comparative analysis of multivariate regression techniques, which included principal component regression (PCR), PLSR, ridge regression, least absolute shrinkage and selection operator (Lasso), and least squares version of support vector machines (LS-SVM) on NIR spectral data of equine articular cartilage. Additionally, we evaluated the effect of variable selection methods, including Monte Carlo uninformative variable elimination (MC-UVE), competitive adaptive reweighted sampling (CARS), variable combination population analysis (VCPA), backward interval PLS (BiPLS), genetic algorithm (GA), and jackknife, on the performance of the optimal regression technique. The PLSR technique was found as an optimal regression tool (R2Tissue thickness = 75.6%, R2Dynamic modulus = 64.9%) for cartilage NIR data; variable selection methods simplified the prediction models enabling the use of lesser number of regression components. However, the improvements in model performance with variable selection methods were found to be statistically insignificant. Thus, the PLSR technique is recommended as the regression tool for multivariate analysis for prediction of articular cartilage properties from its NIR spectra.

Prediction of mineral contents in sugarcane cultivated under saline conditions based on stalk scanning by Vis/NIR spectral reflectance

Mineral contents in the sugarcane juice are important qualitative variables to be used in the development of new products and to establish physicochemical criteria for food processing. The objective of this study was to estimate the mineral contents (K+, Na+, and Mg2+) in sugarcane stalk samples by visible/near infrared (Vis/NIR) spectral reflectance measurements and multivariate data analysis. Four sugarcane varieties were cultivated under different saline conditions in a greenhouse. The spectral reflectance of stalk surface was measured with a portable spectrometer. The juice mineral compositions were determined by a reference laboratory method. Partial least squares regression (PLSR) was used for calibrating the estimation models. The most abundant mineral in the sugarcane juice was K+ (310.01–561.48 mg 100 ml−1) as compared to Na+ (4.63–122.96 mg 100 ml−1) and Mg2+ (6.93–18.78 mg 100 ml−1). During the calibration process, the proposed models presented low root mean square errors (RMSE) for calibration (62.46, 25.70, 2.36, and 0.05 mg 100 ml−1), and for cross-validation (90.08, 30.86, 3.79, and 0.06 mg 100 ml−1) for K+, Na+, Mg2+, and Na+/K+ ratio, respectively. Results were also satisfactory for the external validation with low root mean square error (27.30, 10.69, 0.57, and 0.03 mg 100 ml−1), and mean bias error (−4.45, 6.57, 0.07, 0.01 mg 100 ml−1), as well as high coefficients of determination (0.78, 0.89, 0.93, and 0.74), for K+, Na+, Mg2+, and Na+/K+ ratio, respectively. Spectrometry combined to the PLSR technique resulted in an efficient, quick, and non-destructive method for evaluating sugarcane mineral contents.

Prediction of peroxide value in omega-3 rich microalgae oil by ATR-FTIR spectroscopy combined with chemometrics.

Our work explored, for the first time, monitoring peroxide value (PV) of omega-3 rich algae oil using ATR-FTIR spectroscopic technique. The PV of the developed method was compared by that obtained by standard method of Association of Official Analytical Chemists (AOAC). In this study, peak area integration (PAI), Partial Least Squares Regression (PLSR), and Principal Component Regression (PCR) were used as the calibration techniques. PV obtained by the AOAC method and by FTIR-ATR technique were well correlated considering the peak area related to trans double bonds and chemometrics techniques of PLSR and PCR. Calibration model was established using the band with a peak point at 966cm-1 (990-940cm-1) related to CH out of plane deformation vibration of trans double bond. Algae oil oxidation could be successfully quantified using PAI, PLSR and PCR techniques. Additionally, hierarchical cluster analysis was performed and significant discrimination was observed coherently with oxidation process.

Detection of Broken Kernels Content in Bulk Wheat Samples Using Near-Infrared Hyperspectral Imaging

Broken kernels in stored wheat are contaminants, and these reduce the quality of wheat and increase risk of spoilage. The present study was conducted to predict the amount of broken kernels in bulk Canada Western Red Spring wheat samples using a near-infrared hyperspectral imaging system. The hyperspectral images of bulk wheat samples with different levels of broken kernels (0, 3, 6, 9, 12, and 15 %) were acquired in the 1000–1600 nm wavelength range at a 10-nm interval. The reflectance spectra acquired from these samples were used to develop regression models using principal component regression (PCR) and partial least-squares regression (PLSR) techniques. A tenfold cross-validation technique was used for determining the optimal number of components required to develop these regression models. Among the two regression techniques used, the PLSR technique [with mean square errors of predictions (MSEP), standard error of cross-validation (SECV), and correlation coefficient of 0.483, 0.70, and 0.94, respectively] performed better than the PCR technique (with MSEP, SECV, and correlation coefficient of 0.74, 0.86, and 0.88, respectively). The classification models were developed using the linear discriminant analysis (LDA) and quadratic discriminant analysis (QDA) statistical techniques. Both classifiers accurately classified (100.0 ± 0.0 %) the uncontaminated wheat sample from the higher contamination level (>6 %) wheat samples. Both classifiers also gave similar classification results (QDA 89.8 ± 2.6 % and LDA 87.7 ± 1.6 %) when multi-way classification was performed to classify uncontaminated sample from the contaminated samples.

Recursive variable selection to update near-infrared spectroscopy model for the determination of soil nitrogen and organic carbon

Variable selection is widely accepted as an important step in quantitative analysis of near-infrared (NIR) spectroscopy. However, the variables preselected based on the calibration set might not be representative of the effective variables in future prediction process due to the large variability among soil sample sets. In this work, variable-updating methods (i.e., update both the model coefficients and effective variables in the prediction process) have been applied to support the robustness of the calibration model when it used to predict heterogeneous samples. Partial least squares regression (PLSR), recursive PLSR (RPLSR), and three variable-updating methods, namely variable importance in the projection combined with PLSR (VIP-PLSR), VIP-RPLSR, and uninformative variable elimination combined with PLSR (UVE-PLSR) were used to construct calibration models for the prediction of soil nitrogen (N) and organic carbon (OC) based on NIR spectroscopy. The entire data set was split into calibration set and prediction set according to soil type. The model VIP-RPLSR achieved the optimal performance for soil N and OC. The values of residual prediction deviation (RPD) were 2.9 and 2.8 for N and OC respectively. The results indicated that VIP-RPLSR was able to learn the information from the latest samples by adapting both model coefficients and effective variables at every sample interval. The proposed method VIP-RPLSR has the advantages of wider applicability and better performance for NIR prediction of soil N and OC in comparison with PLSR, RPLSR, VIP-PLSR and UVE-PLSR modeling techniques.

Detecting crop water status in mature olive groves using vegetation spectral measurements

Full spectral measurements (350–2500 nm) at tree canopy and leaf levels and the corresponding leaf water potentials (LWP) were acquired in an olive grove of Sicily, at different hours of the day, during summer season 2011. The main objective of the work was to assess, on the basis of the experimental data-set, two different approaches to detect crop water status in terms of LWP. Specifically, using existing families of Vegetation Indices (VIs) and applying Partial Least Squares Regression (PLSR) were optimised and tested. The results indicated that a satisfactory estimation of LWP at tree canopy and leaf levels can be obtained using vegetation indices based on the near infrared–shortwave infrared (NIR–SWIR) domain requiring, however, a specific optimisation of the corresponding “centre-bands”. At tree canopy level, a good prediction of LWP was obtained by using optimised indices working in the visible domain, like the Normalized Difference Greenness Vegetation Index (NDGI, RMSE = 0.37 and R2 = 0.57), the Green Index (GI, RMSE = 0.53 and R2 = 0.39) and the Moisture Spectral Index (MSI, RMSE = 0.41 and R2 = 0.48). On the other hand, a satisfactory estimation of LWP at leaf level was obtained using indices combining SWIR and NIR wavelengths. The best prediction was specifically found by optimising the MSI (RMSE of 0.72 and R2 = 0.45) and the Normalized Difference Water Index (NDWI, RMSE = 0.75 and R2 = 0.45). Even using the PLSR technique, a remarkable prediction of LWP at both tree canopy and leaf levels was obtained. However, this technique requires the availability of full spectra with high resolution, which can only be obtained with handheld spectroradiometers or hyper-spectral remote sensors.