Linear Discriminant Analysis Research Articles (Page 1)

Electronic noses (e-noses) offer a practical solution for real-time monitoring of ammonia (NH3) in agricultural environments, where NH3 often coexists with interfering gases such as CO2, CH4, and H2S. However, semiconductor-based gas sensors commonly used in e-nose systems suffer from inherent cross-sensitivity, which reduces measurement accuracy. This study investigates the cross-sensitivity of NH3 detection and introduces a mitigation strategy through convolutional neural networks (CNNs) for sensor data fusion. Experimental results show that WO2-based sensors exhibit strong NH3 selectivity, with response ratios of 7.3:1 against CH4 and 17.8:1 against H2S. Density functional theory (DFT) analysis confirmed that the WO3 sensor exhibited strongest NH3 binding energy (- 1.45eV), compared to SnO2 (- 1.10eV), explaining the observed selectivity. Measurement uncertainties (± 8%) were quantified under varying humidity (30-90% RH) and temperature (10-40°C) using a weighted least squares error propagation model. A quasi-2D sensor array improved NH3 classification accuracy to 96.4% (7.2% increase) while reducing concentration errors by 50.8%, as validated by linear discriminant analysis. Long-term stability tests demonstrated that SnO2 sensors maintained a low baseline drift of 0.18%/day over 180days, outperforming CH4 (0.31%/day) and ZnO (0.42%/day) sensors. Furthermore, the CNN model, trained on multi-sensor time-series data, achieved 91.7% accuracy in mixed-gas environments by capturing non-linear response patterns, ensuring reliable NH3 quantification despite interferents. These findings highlight the promise of CNN-enhanced e-nose systems for precise NH3 monitoring in complex agricultural settings, addressing key challenges of cross-sensitivity and environmental stability.

Car classification, using different machine learning models with optimization frameworks, is done for evaluation in this work. We used different models, namely, Extra Trees, XGBoost, Gaussian Naive Bayes, K-Nearest Neighbors, Histogram-based Gradient Boosting (Hist Gradient Boosting), and Linear Discriminant Analysis, for classification. The car samples are classified as “very good,” “good,” “acceptable,” and “unacceptable.” Among these, Hist Gradient Boosting has the highest value for precision, accuracy, recall, and F1 score. We further tuned this model using Evolutionary Strategies, Evolutionary Programming, Covariance Matrix Adaptation Evolution Strategy, and the Flower Pollination Algorithm. Our outcomes indicate that the Covariance Matrix Adaptation Evolution Strategy and Flower Pollination Algorithm significantly enhanced the performance of the model and outperformed Evolutionary Programming. This work investigates the potential of integrating advanced machine learning models with sophisticated optimization strategies to deliver an effective car evaluation classification process that will be useful in this industry and, perhaps, in many other classification tasks.

This study examined the relationship between maternal preeclampsia (PE) and gut microbiota colonization in preterm infants and analyzed the effects of prenatal Bifidobacterium supplementation. This observational study included 45 preterm infants categorized according to their mothers' exposure status during pregnancy. Group A (healthy controls, n = 15) included infants born to healthy mothers who received no supplementation; Group B (PE+Bifidobacterium, n = 15) included infants whose mothers had PE and received Bifidobacterium supplementation as part of routine clinical management; and Group C (PE only, n = 15) included infants born to mothers with PE who did not receive Bifidobacterium supplementation. All enrolled infants were followed from birth for subsequent analyses. The initial postnatal fecal samples of the infants were collected and analyzed using 16S rRNA gene sequencing. Microbial diversity within the intestinal microbiota was evaluated using alpha diversity (within-sample) and beta diversity (between-sample) analyses. To identify taxon-specific differences among groups, we performed linear discriminant analysis effect size and differential abundance analysis, with statistical significance set at p < 0.05. The functional potential of the gut microbiota was inferred based on Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways via the PICRUSt2 algorithm. Alpha diversity analysis revealed significantly greater microbial diversity in the fecal microbiota of preterm infants born to healthy mothers (Group A) than in those delivered by mothers with PE, regardless of prenatal Bifidobacterium exposure. Taxonomic profiling revealed distinct microbial community structures across groups: Group A exhibited significant enrichment of Bacteroides at all taxonomic levels, along with an elevated abundance of Clostridium at the class and order levels. Group B showed a markedly greater relative abundance of Actinobacteria at the phylum level and Rothia at the genus level, whereas Group C was dominated by Proteobacteria (phylum level) and Streptococcus (genus level). All intergroup differences were statistically significant following Benjamini‒Hochberg correction (q < 0.05). A functional analysis of the gut microbiota revealed 53 KEGG pathways with significant overall group differences (p < 0.05), among which 23 pathways were significantly different in at least two groups (q < 0.05). Notably, the activity of the LPS biosynthesis pathway was significantly upregulated in Group C compared with Group A (q = 0.001). Although LPS biosynthesis activity was reduced in Group B relative to Group C (q = 0.018), it remained elevated compared to Group A (q = 0.001), suggesting incomplete mitigation of endotoxin risk. Additionally, glycolytic activity was significantly impaired in Group C relative to Group A (q = 0.003) but was partially restored in Group B compared to Group C (q = 0.022). Maternal PE impaired early-life gut microbiota establishment in preterm infants, manifesting in reduced microbial diversity, enrichment of pathogenic Proteobacteria and Streptococcus, and consequent functional dysbiosis characterized by elevated endotoxin biosynthesis potential and compromised energy metabolism. Although prenatal supplementation with Bifidobacterium partially restored the microbial compositional balance, promoting beneficial bacteria, reducing LPS synthesis activity, and partially improving glycolytic function, it failed to fully reverse endotoxin-related risks, indicating the need to develop more effective microbiota-targeted strategies to comprehensively optimize metabolic and immune homeostasis.

In the present study, an intergeneric cross was conducted between Opsariichthys bidens (♀) and Zacco acanthogenys (♂);the morphological characteristics of the hybrid offspring generation (F1) were observed and recorded at various periods of the development of the embryos and young larvae; and comparative analyses of the external morphological data of the adult fish and the parameters of their external shape framework were carried out. Results demonstrated that hybrid fertilized eggs maintained at (25 ± 1)°C water temperature completed hatching within 42 h 47 min post-fertilization, requiring an accumulated temperature of 1026.72°C·h. Embryonic development consisted of 7 distinct phases subdivided into 28 developmental stages. Newly hatched larvae exhibited a mean total length (TL) of 5.23 ± 0.06 mm, with complete yolk-sac absorption occurring by 7 days post-hatching (dph). Early developmental growth patterns (0-30 dph) conformed to the quadratic equation: y = 0.02096x2 + 0.23443x + 5.6249 (R2 = 0.9559), where y represents TL (mm) and x denotes days post-hatching. Morphological analysis revealed that meristic traits of hybrid progeny predominantly aligned with the maternal lineage, whereas morphometric ratios exhibited intermediate values between parental species. Principal component analysis and cluster analysis of morphometric proportions and framework parameters demonstrated closer similarity to maternal characteristics. Linear discriminant analysis confirmed distinct phenotypic differentiation among the three groups (parental species and hybrids).

In the agri-food supply chain, geographical traceability of milk is essential, quality components being strictly related to the environment and to specific farming. The aim of this study was to test different machine learning approaches for tracing the origin of milk from infra-red spectra in two economically important Italian dairy sheep breeds: Sarda and Valle del Belice. A total of 905 milk samples were collected from Animal farmed in different altimetric zones: plain (≤ 200 metres), hills (350–450), and mountains (> 650 metres). The datasets were divided in training (90%) and testing (10%). Seven models were tested: linear discriminant analysis (LDA), stochastic gradient boosting machine (GBM), support vector machines (SVM), recursive partitioning robust tree (RPART), random forest (RF), K-Nearest neighbours (KNN). The models were calibrated in the training dataset and then used to predict the zone in the testing data set. The procedure was repeated one thousand times by randomly selecting the samples, in the whole dataset or separately by breed. Very different computing time were found, moving from few seconds (LDA and RPART) to more than 20 min (RF). Considering the two breeds together, the LDA model showed the highest accuracy (0.98). The model performances changed according to the breed: GBM was the best model for Sarda, whereas LDA was the best for Valle del Belice. In both breeds, the lowest accuracy was observed for the hill group. The results suggested that this approach can be promising to routinely classify the origin of milk samples using midinfrared spectroscopy.

This study evaluated the behaviour of Linear Discriminant Analysis (LDA) in an imbalanced dataset where two LDA models were considered: the classical Fisher linear Discriminant Analysis (CFLDA), which is the same as the LDA and the Robust Fisher Linear Discriminant Analysis (RFLDA). The study applied the Monte Carlo simulation to investigate the comparative performance of both classifiers. Also, an investigation was done on both classifiers using practical datasets. The violation of the assumptions of the LDA model was observed, and the satisfaction of the central limit theorem was observed in the performance of the classifiers. The imbalance data concept associated with the practical data and the impact of data balancing using the Mean Variance Cloning Techniques (MVCT) was also demonstrated. The analysis demonstrated the comparative performance of the classifiers and also indicated the weaknesses of both classifiers. The results demonstrated that the RFLDA performance when faced with contamination and alteration in both training and validation samples is not affected, but will perform better than the CFLDA if the assumptions (normality and Homoscedasticity) are violated, as the RFLDA can resist noise. In general, the result showed that RFLDA is not susceptible to contamination and alteration, but the CFLDA was shown to perform well on an imbalanced sample size, thereby validating the Concept of data dependency and central limit theory.

Abstract Fast-growing wood that imitates colours develops colour differences over time. This shortens its service life. To solve the problems of short colour retention time and insufficient accuracy of colour imitation. In this study, a prediction method for Mongolian scots pine dyeing formulation based on 1D convolutional neural networks (CNN), Bidirectional gated recirculating unit network, linear discriminant analysis and attention mechanism is proposed. First, a large database was established through experiments involving the dyeing of Mongolian scots pine wood chips and multispectral measurements. Then, linear discriminant analysis was used for feature extraction, classification and dimension reduction of multispectral information to reduce the size of data input. Then, the data were fed into a one-dimensional convolutional neural network for feature re-extraction and recipe prediction by a two-way gated recurrent unit network, while an attention mechanism was introduced to highlight key spectral segments and improve the efficiency of the network model. The evaluation results show that the segments and improve the efficiency of the network model. The evaluation results show that the multispectral data input significantly improves the colour difference problem of colour imitation over time, and the present model significantly improves the accuracy of colour imitation. According to the International Commission on Illumination (CIE) 2000 color difference formula (CIEDE2000), the model achieved a minor-difference grade of 99.39% and a no-difference grade of 85.63%, with a coefficient of determination ( R 2 ) of 0.95.

Abstract Insects and microbes form diverse symbiotic associations that are fundamental to ecosystem functioning and the adaptive success of hosts. For instance, some Burkholderia sensu lato (s.l.) bacteria are facultative symbionts in stinkbug taxa, acquired from the environmental soil in each generation. Based on large‐scale sampling and deep sequencing data, we hope to reveal the diversity of Burkholderia in stinkbug species across various localities, determine the factors influencing the Burkholderia communities within stinkbugs and explore the strain concordance pattern between stinkbugs and Burkholderia in the soil. The Burkholderiaceae sequences accounted for an average of 75.1% of all bacterial sequences in the 215 stinkbug samples. Among them, Caballeronia , Paraburkholderia and Burkholderia sensu stricto (s.s.) represented more than 99.9% of the total number of Burkholderiaceae. The linear discriminant analysis effect size (LEfSe analysis) revealed hierarchical specificity patterns in Burkholderia ‐stinkbug associations across host taxonomic levels. Beta diversity analysis revealed that both the geographical location and classification status of stinkbugs significantly affect Burkholderia communities; however, the proportion of variance explained was consistently low. Several low‐abundance strains of soil Burkholderia appeared frequently in stinkbug individuals, suggesting that differences in competitiveness might exist between strains. Our results also suggested that the colonisation of Burkholderia in stinkbugs is, to some extent, strain density‐dependent when the environment contains multiple closely related Burkholderia strains. These findings deepened our understanding of the symbiosis between stinkbugs and Burkholderia in their natural state.

Rapid, sensitive, and specific detection of pathogenic Escherichia coli serotypes is crucial for food safety and public health. Here, we present a surface-enhanced Raman scattering (SERS) platform utilizing highly ordered silver nanorod (AgNR) arrays functionalized with vancomycin for efficient and selective bacterial capture. The system enables multiplexed, high-throughput analysis using a portable Raman spectrometer, achieving direct molecular fingerprinting of seven clinically relevant E. coli serotypes. Systematic optimization of AgNR length and vancomycin coating maximized SERS enhancement and capture efficiency. Advanced data analysis with linear discriminant analysis (LDA) provided robust discrimination among all serotypes and concentrations, achieving up to 100% classification accuracy in single-concentration models and an overall accuracy of 98.41% when all concentrations and serotypes were evaluated jointly. This integrated SERS approach demonstrates significant promise for rapid, on-site bacterial diagnostics and quantitative pathogen monitoring, paving the way for practical applications in food safety and clinical microbiology.

Exhaled breath analysis is emerging as one of the most promising non-invasive strategies for the early detection of life-threatening diseases, especially lung cancer, where rapid and reliable diagnosis remains a major clinical challenge. In this study, we designed and optimized an electronic nose (e-nose) platform composed of quantum resistive vapor sensors (vQRSs) engineered by polymer-carbon nanotube nanocomposites via spray layer-by-layer assembly. Each sensor was tailored through specific polymer functionalization to tune selectivity and enhance sensitivity toward volatile organic compounds (VOCs) of medical relevance. The sensor array, combined with linear discriminant analysis (LDA), demonstrated the ability to accurately discriminate between cancer-related biomarkers in synthetic blends, even when present at trace concentrations within complex volatile backgrounds. Beyond artificial mixtures, the system successfully distinguished real exhaled breath samples collected under challenging conditions, including before and after smoking and alcohol consumption. These results not only validate the robustness and reproducibility of the vQRS-based array but also highlight its potential as a versatile diagnostic tool. Overall, this work underscores the relevance of nanocomposite chemo-resistive arrays for breathomics and paves the way for their integration into future portable e-nose devices dedicated to telemedicine, continuous monitoring, and early-stage disease diagnosis.

Background: Accurate reconstruction of the orbit after trauma or oncological resection requires reliable anatomical references. In unilateral cases, the contralateral orbit can guide repair, but bilateral injuries or pathologies remove this option. To address this problem, we developed a new morphological classification of orbits based on three linear dimensions. Methods: A total of 499 orbits from patients of Caucasian descent (age 8–88 years) were analyzed using three-dimensional models generated from cone-beam and fan-beam CT scans. Orbital depth (D), height (H), and width (W) were measured, and proportional indices were calculated. K-means clustering (k = 3) identified recurring morphotypes, validated by linear discriminant analysis (LDA) and supported by ANOVA, Kruskal–Wallis, and correlation tests (age and sex). Results: Three morphotypes were identified: Tall &amp; Broad (type A, 33.5%), Deep &amp; Broad (type B, 30.2%), and Compact (type C, 36.2%). All dimensions differed significantly between groups (ANOVA, p &lt; 1 × 10−16; η2 = 0.40–0.51). Male orbits were significantly deeper and wider than female ones (p &lt; 0.001). LDA demonstrated excellent separation with 97.5% accuracy. A simplified decision algorithm achieved 82.1% classification accuracy. In situations where only orbital depth could be measured, an alternative cut-off-based method reached 61.5% accuracy, with type B and C better distinguished than type A. Conclusions: The proposed classification provides a reproducible framework for describing orbital morphology. It may serve as a reference in cases where local anatomy is disrupted or the contralateral orbit is unavailable. Even millimeter-scale differences in orbital dimensions may correspond to clinically relevant changes in orbital volume and globe position, underlining the potential usefulness of this system in surgical planning.

Identification of common ions in water is essential for the evaluation of beverage quality and utilization of water resources. However, most current approaches focus on the determination of water safety, and those for assessment of water quality are very rare. In this work, we took beer as an example and proposed an optical sensor array based on a series of supramolecular probes for the identification of six key brewing ions (KBIs) (Na+, Mg2+, Ca2+, HCO3-, SO42-, and Cl-) in brewing water. Six sensing units were designed and screened according to the different natures of each KBI, such as pKa, electrified characteristics, coordination properties, etc. Spectroscopic studies indicated that KBIs could regulate the assembly and disassembly of supramolecular probes through noncovalent interactions (electrostatic, hydrophobic, π-π stacking, etc.), resulting in cross-reactive optical signals. In combination with linear discriminant analysis (LDA), a fingerprint database of brewing water quality was constructed for the first time with good accuracy, simplicity, and low cost simultaneously. Based on this database, different concentrations of single ions and their binary mixtures were distinguished with 100% identification accuracy. It can be further applied for rapid matching of water characteristics and beer styles and provides a strategy for precision brewing and targeted water resource development.

Abstract. The ecology of benthic foraminifera in tidal flats has been extensively studied at local scales, but its seasonal and spatial dynamics across broader intertidal gradients remain poorly understood. This study investigates seasonal variations in the composition and structure of living benthic foraminiferal assemblages across contrasting intertidal habitats in the eastern English Channel. We analyzed 256 surface sediment samples (192 for foraminifera and 64 for sediment properties) collected over four seasons at various sites in the Hauts-de-France region (northern France) along the eastern English Channel. Multivariate analyses revealed significant seasonal changes in the assemblage structure. Linear discriminant analysis identified two dominant seasonal groups. Opportunistic taxa, such as Haynesina germanica and Ammonia confertitesta, dominate during colder seasons, while more diverse and thermophilic species including Elphidium selseyense, Quinqueloculina dimidiata, and Trochammina inflata characterize warmer months. These findings provide new insights into the phenology of benthic foraminifera and contribute to a better understanding of seasonal ecological processes in temperate intertidal ecosystems.

This study investigates the taxonomic status of Salvia alba J.R.I. Wood and S. personata Epling, two previously synonymized species belonging to the subgenus Calosphace, through an integrative approach combining traditional morphology, geometric morphometrics of flower shape and phylogenetic analysis. Morphologically, living specimens of S. alba are characterized by larger, pure white corollas with wider throats and a ventricose tube, while S. personata has blue and narrower corollas, with a straight tube. Although some morphological measurements overlap in herbarium specimens, the relative length of the corolla tube to the flower-bearing calyx consistently remains a distinguishing feature: in S. alba, the corolla tube exceeds the calyx length, unlike in S. personata. No significant differences were observed in leaf morphology. The geometric morphometric analysis further supported the distinction between the two taxa and linear discriminant analysis based on flower shape variables classified the species with 100% accuracy. Molecular phylogenetics based on the ITS nuclear region placed S. alba and S. personata within different subclades of the Angulatae clade. Based on this comprehensive evidence, we conclude that S. alba should be reinstated as a valid species, with the length ratio between corolla and calyx a key diagnostic character, especially in preserved specimens, while corolla colour, size and shape are consistent and distinctive characters in living specimens.

Prediction of T2/T3 Staging in Patients with Volume-Equivalent Esophageal Squamous Cell Carcinoma on the Basis of PET/CT Radiomics.

Sensor array based on Cu-based metal-organic frameworks nanozymes for rapid identification of phenolic pollutants.

Machine learning-assisted construction of a lignin carbon dots sensor array for detecting food colorants.

A machine learning assisted approach to classify rose species and varieties with laser induced breakdown spectroscopy.

A data fusion system based on attenuated total reflectance mid-infrared spectroscopy and colorimetry combined with chemometrics for monitoring the fermentation process of Candida utilis.

Sensitive discrimination of hazardous explosives by a sensor array based on siloles with aggregate-induced emission.