Locality preserving partial least squares discriminant analysis for face recognition

In the present study, boosting has been combined with partial least‐squares discriminant analysis (PLS‐DA) to develop a new pattern recognition method called boosting partial least‐squares discriminant analysis (BPLS‐DA). BPLS‐DA is implemented by firstly constructing a series of PLS‐DA models on the various weighted versions of the original calibration set and then combining the predictions from the constructed PLS‐DA models to obtain the integrative results by weighted majority vote. Coupled with near infrared (NIR) spectroscopy, BPLS‐DA has been applied to discriminate different kinds of tea varieties. As comparisons to BPLS‐DA, the conventional principal component analysis, linear discriminant analysis (LDA), and PLS‐DA have also been investigated. Experimental results have shown that the inter‐variety difference can be accurately and rapidly distinguished via NIR spectroscopy coupled with BPLS‐DA. Moreover, the introduction of boosting drastically enhances the performance of an individual PLS‐DA, and BPLS‐DA is a well‐performed pattern recognition technique superior to LDA. Copyright © 2012 John Wiley & Sons, Ltd.

Partial least squares discriminant analysis (PLS-DA) is gaining popularity as a discriminant feature extraction tool. It is often viewed as a “supervised” version of the principle component analysis (PCA) where dimensionality reduction is achieved with full awareness of the class labels. It is well known that as a dimensionality reduction tool, PCA only captures the global geometric structure of a data set and this characteristics of PCA is inherited in PLS-DA. In this paper we propose a locality preserving PLS-DA (LPPLS-DA) in an effort to enhance the discriminant capabilities of PLS-DA. Our proposed LPPLS-DA incorporates neighborhood information of the data by means of a similarity matrix. We investigate the use of LPPLS-DA in spoken letter recognition. Finding relevant features for the classification of spoken letters is affected by high dimensionality of the data and the sound (phonetic) similarities between the letters. To address this problem, we used LPPLS-DA to extract relevant features of the data and discrimination is automatically achieved using the LPPLS-DA method. The effectiveness of LPPLS-DA is demonstrated by experimental results on the ISOLET (Isolated Letter Speech Recognition) data set. The classification accuracies in the experiments are measured using the nearest neighbor classifier. The LPPLS-DA method is shown to give higher accuracies than PLS-DA as well as several other feature extraction methods.

Due to the existence of Lingzhi adulteration, there is a growing demand for species classification of medicinal mushrooms by various techniques. The objective of this study was to explore a rapid and reliable way to distinguish between different Lingzhi species and compare the influence of data pretreatment methods on the recognition results. To this end, 120 fresh fruiting bodies of Lingzhi were collected, and all of them were analyzed by attenuated total reflection-Fourier transform infrared spectroscopy (ATR-FTIR). Random forest (RF), support vector machine (SVM) and partial least squares discriminant analysis (PLS-DA) classification models were established for raw and pretreated second derivative (SD) spectral matrices to authenticate different Lingzhi species. The results of multivariate statistical analysis indicated that the SD preprocessing method displayed a higher classification ability, which may be attributed to the analysis of powder samples that requires removal of overlapping peaks and baseline shifts. Compared with RF, the results of the SVM and PLS-DA methods were more satisfying, and their accuracies for the test set were both 100%. Among SVM and PLS-DA, the training set and test set accuracy of PLS-DA were both 100%. In conclusion, ATR-FTIR spectroscopy data pretreated by SD combined with PLS-DA is a simple, rapid, non-destructive and relatively inexpensive method to discriminate between mushroom species and provide a good reference to quality assessment.

A comparative study on classification of edible vegetable oils by infrared, near infrared and fluorescence spectroscopy combined with chemometrics

Aflatoxin contamination can occur in a wide variety of agricultural products pre- and post-harvest, posing potential severe health hazards to human and livestock. However, current methods for detecting aflatoxins are generally based on wet chemical analyses, which are time-consuming, destructive to test samples and require skilled personnel to perform, making them impossible for large-scale non-destructive screening and on-site detection. In this study, we utilized the visible/near-infrared (Vis/NIR) spectroscopy over the spectral range of 400-2500 nm to detect contamination of shelled commercial peanut kernels with the predominant aflatoxin B1 (AFB1). Our results indicated the usefulness of Vis/NIR spectroscopy combined with the chemometric techniques of partial least squares discriminant analysis (PLS-DA) and least squares support vector machine (LS-SVM) in identifying the AFB1 contamination of peanut kernels. Both PLS-DA and LS-SVM methods provided satisfactory classification results using the full spectral information over the ranges of 410-1070 (I), 1120-2470 nm (II) and I+II. Based on the classification threshold of 20 ppb, the best PLS-DA prediction results using the full spectra yielded the average accuracy of 87.9% and overall accuracy of 88.6%. With 100 ppb as the classification threshold, the best PLS-DA model using the full spectra achieved the average accuracy of 94.0% and overall accuracy of 91.4%. Using the full spectra, the best average accuracies recorded by LS-SVM were 90.9% and 98.0%, with the classification thresholds of 20 and 100 ppb, respectively. Correspondingly, the best overall accuracies by LS-SVM were 90.0% and 97.1%. In addition, the simplified models of CARS-PLS-DA and CARS-LS-SVM also demonstrated good prediction capability in identifying the AFB1 contamination from peanut surface. Based on both classification thresholds of 20 and 100 ppb, the best CARS-PLS-DA and CARS-LS-SVM prediction results were ≥ 90.0% in both average accuracy and overall accuracy. Most importantly, the computation complexity and the employed data dimensionality were significantly reduced by using the simplified models.

Endometrial cancer (EC) is one of the most common gynecologic neoplasms in developed countries but lacks screening biomarkers. We aim to identify and validate metabolomic biomarkers in cervicovaginal fluid (CVF) for detecting EC through nuclear magnetic resonance (NMR) spectroscopy. We screened 100 women with suspicion of EC and benign gynecological conditions, and randomized them into the training and independent testing datasets using a 5:1 study design. CVF samples were analyzed using a 600-MHz NMR spectrometer equipped with a cryoprobe. Four machine learning algorithms-support vector machine (SVM), partial least squares discriminant analysis (PLS-DA), random forest (RF), and logistic regression (LR), were applied to develop the model for identifying metabolomic biomarkers in cervicovaginal fluid for EC detection. A total of 54 women were eligible for the final analysis, with 21 EC and 33 non-EC. From 29 identified metabolites in cervicovaginal fluid samples, the top-ranking metabolites chosen through SVM, RF and PLS-DA which existed in independent metabolic pathways, i.e. phosphocholine, malate, and asparagine, were selected to build the prediction model. The SVM, PLS-DA, RF, and LR methods all yielded area under the curve values between 0.88 and 0.92 in the training dataset. In the testing dataset, the SVM and RF methods yielded the highest accuracy of 0.78 and the specificity of 0.75 and 0.80, respectively. Phosphocholine, asparagine, and malate from cervicovaginal fluid, which were identified and independently validated through models built using machine learning algorithms, are promising metabolomic biomarkers for the detection of EC using NMR spectroscopy.

The photoreversible metal-ion complexation behavior of nitroquinolinospiropyranindoline (NQSP) was studied in combination with the selective identification of six different transition-metal ions, Zn(2+), Co(2+), Hg(2+), Cu(2+), Cd(2+), and Ni(2+), in single- and binary-component mixtures via partial least squares discriminant (PLSD) analysis. The plasticizer, dicapryl phthalate, was chosen as the support medium for this study on the basis of (1) its enhancement of the photoreversibility and hypsochromic shifts seen in metal complexation and (2) its potential application to supported liquid membranes for eventual sensor applications. Complexation of divalent transition-metal ions by NQSP in dicapryl phthalate produced variable hypsochromic shifts in the absorption spectra (30-60 nm), requiring chemometric techniques in order to overcome the spectral overlaps. PLSD analysis was used to build classification analysis models to differentiate between the six divalent transition-metal ions. The feasibility of performing mixture analysis was studied using the concept of net analyte signal prior to experimental verification. Single- and binary-component mixtures of metals were identified with 100 and 97.4% accuracy, respectively, which included no false positives in either the training or prediction sets.

Currently, reliable biomarkers that can be used to distinguish rheumatoid arthritis (RA) from other inflammatory diseases are unavailable. To find possible distinctive metabolic patterns and biomarker candidates for RA, we performed global metabolite profiling of synovial fluid samples. Synovial fluid samples from 38 patients with RA, ankylosing spondylitis, Behçet's disease, and gout were analyzed by gas chromatography/time-of-flight mass spectrometry (GC/TOF MS). Orthogonal partial least-squares discriminant and hierarchical clustering analyses were performed for the discrimination of RA and non-RA groups. Variable importance for projection values were determined, and the Wilcoxon-Mann-Whitney test and the breakdown and one-way analysis of variance were conducted to identify potential biomarkers for RA. A total of 105 metabolites were identified from synovial fluid samples. The score plot of orthogonal partial least squares discriminant analysis showed significant discrimination between the RA and non-RA groups. The 20 metabolites, including citrulline, succinate, glutamine, octadecanol, isopalmitic acid, and glycerol, were identified as potential biomarkers for RA. These metabolites were found to be associated with the urea and TCA cycles as well as fatty acid and amino acid metabolism. The metabolomic analysis results demonstrated that global metabolite profiling by GC/TOF MS might be a useful tool for the effective diagnosis and further understanding of RA.

The accurate and non-destructive assessment of fruit ripeness is essential for post-harvest sorting and quality management. This study evaluated a meta-inspired classification framework integrating partial least squares discriminant analysis (PLS-DA) with support vector machines (SVMs) trained on latent variables (sSVM) or on class probabilities (pSVM) derived from multiple PLS-DA components. Two kiwiberry varieties, ‘Geneva’ and ‘Weiki’, were analyzed using variety-specific and combined datasets. Performance was assessed in calibration and prediction using accuracy, F05, Cohen’s kappa, precision, sensitivity, specificity, and likelihood ratios. Conventional PLS-DA provided reasonably good classification, but pSVM models, particularly those with an RBF kernel (pSVM_R), consistently outperformed other approaches and ensured higher stability across all datasets. Unlike sSVMs, which were prone to overfitting, pSVM_R models achieved the highest accuracy of 92.4–96.9%, Cohen’s kappa of 84.8–93.9%, and precision of 89.1–94.2%, clearly surpassing both score-based SVM and PLS-DA. Contrasting tendencies were observed between cultivars: ‘Geneva’ models improved during prediction, while ‘Weiki’ models declined, especially in specificity. Combined datasets provided greater stability but slightly reduced peak performance than single-variety models. These findings highlight the value of probability-enriched stacking models for non-invasive ripeness discrimination, suggesting that adaptive or hybrid strategies may further enhance generalization across diverse cultivars.

Untargeted metabolomics is a well-established technique and a powerful tool to find potential plasma biomarkers for early diagnosing hereditary transthyretin amyloidosis. Hereditary transthyretin amyloidosis (ATTRv) is a disabling and fatal disease with different clinical features such as polyneuropathy, cardiomyopathy, different gastrointestinal symptoms and renal failure. Plasma specimens collected from 27 patients with ATTRv (ATTRV30M), 26 asymptomatic TTRV30M carriers and 26 control individuals were subjected to gas chromatography (GC)- and liquid chromatography (LC)-mass spectrometry (MS)-based metabolomics analysis. Partial least squares discriminant and univariate analysis was used to analyse the data. The models constructed by Partial least squares-discriminant analysis (PLS-DA) could clearly discriminate ATTRV30M patients from controls and asymptomatic TTRV30M carriers. In total, 24 plasma metabolites (VIP > 1.0 and p < .05) were significantly altered in ATTRV30M patient group (6 increased and 18 decreased). Eleven of these distinguished the ATTRV30M group from both controls and TTRV30M carriers. Plasma metabolomics analysis revealed marked changes in several pathways in patients with ATTRV30M amyloidosis. Statistical analysis identified a panel of biomarkers that could effectively separate controls/TTRV30M carriers from ATTRV30M patients. These biomarkers can potentially be used to diagnose patients at an early stage of the disease.

Near infrared (NIR) spectroscopy with chemometrics was used for rapid source identification and component quantification of Arnebiae Radix. The modeling parameters were determined using the processing trajectory method. The quantitative model of target components was established using the model fusion method. The root mean square error of calibration (RMSEC), root mean square error of validation (RMSEP), determination coefficient of calibration (Rcal 2 ) and validation (Rpre 2 ), and the ratio of standard error of prediction to standard deviation (RPD) were used to evaluate the constructed model. The RMSEC, Rcal 2 , RMSEP, Rpre 2 , and RPD of β-β'-dimethylacryloacanin partial least square (PLS) model established using the step-by-step strategy were 1.3004 mg/g, 0.9453, 1.3833 mg/g, 0.9544, and 4.6650, respectively. The RMSEC, Rcal 2, RMSEP, Rpre 2 , and RPD of total pigment PLS model were 4.5095 mg/g, 0.9533, 4.5679 mg/g, 0.9534, and 3.0509, respectively. The RMSEC, Rcal 2 , RMSEP, Rpre 2 , and RPD of β-β'-dimethylacryloacanin Bagging-PLS model established using the process trajectory were 1.6126 mg/g, 0.9159, 1.2254 mg/g, 0.9649, and 5.2662, respectively. The RMSEC, Rcal 2 , RMSEP, Rpre 2 , and RPD of the total pigment Bagging-PLS model were 4.3499 mg/g, 0.9565, 4.4662 mg/g, 0.9116, and 3.1205, respectively. The accuracy of quantitative model may be improved using the process trajectory and model fusion method. Partial least squares discriminant analysis (PLS-DA) was used to establish the discriminant models of different sources of Arnebiae Radix. The predicted accuracy rate was 100%. The method is simple, rapid, and demonstrated for the discrimination and quantitative analysis of the main active ingredients of Arnebiae Radix.

Spectroscopic methods have already been used as effective tools in several studies involving the detection of cancer. Fourier transform infrared spectroscopy (FTIR) has already been applied in the discrimination of cancer cells and tissues or blood of patients with the disease, observing that this technique requires the use of chemometric algorithms to obtain such results. The aim of this study was to employ a partial least squares discriminant analysis (PLS-DA) with FTIR data in the discrimination of plasma samples from patients with colorectal cancer (RCC) and healthy individuals. Multivariate analysis was performed using PLS-DA of the sample triplicates (n=90) with different types of processing. The best PLS-DA condition was obtained using the 1st derivative, 1 orthogonal signal correction (OSC) and no pre-processing. With 1 factor only, the model presented a mean square error of cross-validation (RMSECV) of 0.0004 and coefficient of determination (r^2) of 1.0000. The accuracy, precision and sensitivity of the model were 100%.

The problems of adulteration and mislabeling are very common in the food industry. Laser-induced breakdown spectroscopy (LIBS) coupled with chemometric methods has many intrinsic advantages on adulteration analysis of various materials. In this work, several chemometric algorithms, i.e., principal component analysis (PCA), decision tree (DT), random forest (RF), partial least squares discriminant analysis (PLS-DA), linear discriminant analysis (LDA), and support vector machine (SVM), were carried out assisted by LIBS technology to study the classification performances on rice geographic origins. A series of samples, including 20 kinds of rice samples from different geographic origins, was detected using LIBS with no pretreatment processes. For data analysis, PCA was employed to reduce the input variables, and to reduce the collinearity of LIBS spectral results as well. The results showed the classification accuracies of the mentioned chemometric algorithms of DT, RF, PLS-DA, LDA, and SVM with 89 input variables of 86.80%, 96.30%, 96.80%, 98.60%, and 99.20%, respectively. At the same time, the operation times of these algorithms were 3.81 s, 54.64 s, 3.63 s, 2.09 s, and 531.01 s, respectively. On the other hand, 30 principal components of input variables were also tested under the same conditions. The classification accuracies for the above algorithms were 81.60%, 98.00%, 95.70%, 98.40%, and 99.20%, respectively. The operation times were 2.01 s, 4.88 s, 3.67 s, 0.36 s, and 308.55 s, respectively. In addition, the five-fold cross-validation classification accuracies with 30 input variables for DT, RF, PLS-DA, LDA, and SVM were 83.75%, 97.95%, 94.75%, 98.35%, and 99.25%, respectively. As a result, LDA was demonstrated to be the best and most efficient tool for rice geographic origin classification assisted by LIBS with high accuracy and analytical speed, which has great potential for rapid identification of adulterated products in agriculture without use of any chemical reagent.

ABSTRACTUltraviolet-visible (UV-Vis) spectral information (190–400 nm) was used to classify coffee samples into either pure peaberry or pure normal coffee classes using two chemometric methods: soft independent modelling of class analogy (SIMCA) and partial least squares discriminant analysis (PLS-DA). The spectral data of peaberry and normal coffee were acquired using a UV-Vis spectrometer (Genesys™ 10S UV-Vis, Thermo Scientific, USA). For both supervised discrimination methods, SIMCA and PLS-DA, all samples were correctly classified into their corresponding classes. The SIMCA model classified all samples accurately (100%) into either the peaberry or normal coffee class, even at a 5% confidence level; however, the PLS-DA model also correctly classified all samples (100%). Investigation of the major wavelengths contributing to the classifications using x-loading weights and loading of latent variables (LVs) indicated wavelengths at 230, 250, 270, 310, and 350 nm were important for determination of the coffee types. These wavelengths were closely related to the absorbance wavelengths of several important chemical components in roasted coffee: caffeine, caffeic acid, and chlorogenic acids (CGA). These results provided the basis for developing a simple and reliable method for peaberry coffee authentication (including more complex samples where peaberry coffee is blended with normal coffees) based on UV-Vis spectra.

P175 Comparative metabolomic and lipidomic analysis of serum samples from patients with seronegative rheumatoid arthritis and psoriatic arthritis