Search Results (2,945)

This study evaluates calibration strategies for predicting soluble solids content (SSC) in Ambrosia apples using a low-cost, portable short-wave near-infrared (SW-NIR) spectrometer (640–1050 nm) under interactance acquisitions. Replicate sampling spectral curves exhibited notable variability due to asymmetric illumination, peel color heterogeneity, and probe contact inconsistencies. Regression models were comparatively built on averaged spectra compared with those trained directly on multiply sampled replicate spectra, applying piecewise Savitzky–Golay smoothing and detrending as pretreatment. Variable selection was performed via uninformative variable elimination (UVE) and backward interval partial least squares (BiPLS). Models calibrated on replicate spectra demonstrated superior generalization to unseen replicate measurements, despite slightly higher cross-validation errors. The BiPLS model on replicate spectra achieved the best predictive performance (mean RMSEP = = 0.677 °Brix, Rp = 0.796, RPD = 1.656), with improved trueness (lower relative absolute bias) and precision (lower relative standard deviation). For comparison, the BiPLS model on averaged spectra yielded a mean RMSEP = 0.899 °Brix, Rp = 0.593, RPD = 1.25; the replicate-spectra strategy thus reduced the RMSEP by 24.7% and increased Rp and RPD accordingly. This suggests that for low-cost NIR instruments, using replicate sampling spectral modeling combined with interval variable selection can provide better prediction performance and achieve the purpose of on-site sorting in food quality analysis. Full article

This study investigated how detection distance affects spectral models for soluble solids content (SSC) and firmness evaluation in Korla fragrant pears and provides a reference for calibrating standardized indoor non-destructive detection equipment. Two hundred visually intact fruit samples at the early-ripening stage were collected from the Korla production area in Xinjiang. An FS-640 multispectral camera system equipped with a VS-SWR fixed-focus industrial lens (16 mm focal length, F1.8 maximum aperture, 1/2-inch sensor format) was used to acquire fruit reflectance spectra at seven vertical lens-to-fruit-surface distances of 90, 100, 110, 120, 130, 140, and 150 cm. A 625-pixel region of interest (ROI) was selected using ENVI at an undamaged equatorial or near-equatorial position of each fruit, and the regional mean spectrum was used as the spectral feature of one fruit sample. The sample-set partitioning based on joint X–Y distances (SPXY) algorithm was used to divide the calibration and prediction sets at a 3:1 ratio after outlier removal via a residual-threshold method. Four preprocessing methods, namely LOESS smoothing, standardization, vector normalization, and Savitzky–Golay (SG) smoothing, were compared. Competitive adaptive reweighted sampling (CARS) was performed with 50 Monte-Carlo sampling runs, a maximum of 30 principal components, and 10-fold cross-validation, yielding 99 characteristic wavelengths. Partial least squares regression (PLSR), support vector regression (SVR), random forest (RF), and artificial neural network (ANN) models were then established using identical input variables and sample partitions. Model performance was evaluated using the coefficient of determination for calibration ($R_c^2$), coefficient of determination for prediction ($R_P^2$), root-mean-square error of calibration (RMSEC), root-mean-square error of prediction (RMSEP), relative prediction deviation (RPD), and ratio of performance to interquartile distance (RPIQ). Under the static laboratory acquisition conditions in this work, the SSC model achieved the best prediction performance at 110 cm with SG smoothing ($R_P^2$) = 0.8949, RPD = 3.0633, RPIQ = 5.8661), whereas the firmness model obtained optimal prediction performance at 140 cm with standardization ($R_P^2$) = 0.7460, RPD = 1.9425, RPIQ = 3.2867). Changes in detection distance altered illumination uniformity, effective reflected signal, photon-scattering paths, and background-noise proportion. These effects may partially explain why the chemical-absorption-dominated SSC index and the tissue-scattering-dominated firmness index responded differently to detection distance. The results provide a reference for setting spectral detection parameters for Korla fragrant pears; however, samples were obtained from only a single producing region, harvest season, and maturity stage, and no independent external validation dataset was used. Therefore, the generalization ability of the developed models needs to be further verified using cross-season and cross-orchard sample sets. Full article

Generative AI shopping assistants are becoming a primary touchpoint for online consumers, yet they often produce convincing but inaccurate content, a phenomenon known as AI hallucination that poses an under-studied risk to consumer trust in e-commerce. This study explains that phenomenon by building and testing a moderated mediation model grounded in Expectation Violation Theory, Epistemic Vigilance Theory, and Algorithmic Trust Repair Theory. A sequential, exploratory mixed-methods design was used: a qualitative phase identified the dimensions and configurational pathways of consumer trust withdrawal using the Gioia methodology and fuzzy-set Qualitative Comparative Analysis, and a subsequent large-scale quantitative phase tested and refined the resulting model across a multi-country European sample using partial least squares structural equation modeling and Necessary Condition Analysis. The results show that exposure to hallucinations triggers expectation violation, activating epistemic vigilance and reducing perceived AI competence; this sequence drives trust recalibration, reflected in lower continued-use and purchase intentions and greater negative word-of-mouth. AI literacy, prior trust, and transparency cues significantly moderate these relationships, and structural trust repair mechanisms, namely retrieval-augmented generation and uncertainty disclosure, prove more effective than purely communicative repair strategies. Theoretically, this study advances a dynamic account of trust recalibration in AI-mediated commerce; practically, it offers concrete guidance for platform design and regulatory policy under the EU AI Act. Full article

Open banking enables customers to provide third parties with financial information, yet this will be successful only when customers are ready to share it. Previous research concentrates on security, trust, and usefulness as direct motivators without paying attention to mechanisms. This research fills this gap by evaluating the effect of Perceived Security Assurance, Trust in Banks and FinTech Providers, and Perceived Usefulness on Willingness to Share Financial Data based on Perceived Data Control, which is based on Privacy Calculus Theory. A combination of cluster and purposive sampling was used to collect data from digital banking users in five regions in Saudi Arabia. There were a total of 384 valid responses that were analysed. SmartPLS 4 was used to run Partial Least Squares Structural Equation Modelling. The results reveal that PSA, TBFP, PUOB, and PDC directly impact WSFD and that PSA, TBFP, and PUOB also have a considerable impact on PDC. The outcomes of the mediation confirm that PDC mediates these relationships partially. Multi-group analysis also shows that these effects are greater amongst users with less experience in FinTech and with younger customers, especially in the pathways that involve perceived data control. The work adds to the theory in two ways. The extension of the Privacy Calculus Theory by adding the perceived control and pointing out the heterogeneity of users in data-sharing behaviour. It also changes the attention from the general adoption intention to the actual data-sharing behaviour in open banking. Practically, the results suggest that to foster customer engagement in open banking ecosystems, it is necessary to improve security, develop trust, prove value, provide user control, and use segment-specific actions. Full article

Existing reviews on AI in tea production are either agriculture-generic or limited to isolated tasks. This review thoroughly compares vision technologies (RGB, hyperspectral, near-infrared, thermal, Light Detection and Ranging (LiDAR), Unmanned Aerial Vehicle (UAV)) and establishes a task-oriented algorithm selection framework for the tea industry. For small-sample or near-linear problems, traditional machine learning (ML) (support vector machine (SVM); partial least squares regression (PLSR)) remains effective. For unstructured field tasks, deep learning achieves superior performance: pest detection accuracy exceeds 97%, tea bud detection reaches 96.8% with RGB images, and hyperspectral imaging predicts nitrogen content with R² > 0.90 and tea polyphenols with R² up to 0.925. Algorithm choice further differentiates by task granularity: lightweight convolutional neural networks (CNNs) balance speed and accuracy for edge deployment at 16 fps; You Only Look Once (YOLO) series detectors enable real-time localization on mobile platforms at 93.1% accuracy, 24 ms per target. No single algorithm dominates all tea tasks; selection is a trade-off among accuracy, speed, data availability, and computational constraints. These findings outline a structured analysis of the challenges and pathways for transitioning computer vision (CV) from laboratory research toward field-deployable tools. Full article

High-fidelity reservoir simulation is too costly for multi-query tasks such as history matching and production optimisation. Existing reduced-order models (ROMs) mitigate this cost but generally lack uncertainty quantification, spatial sparsity, and interpretable mode-to-geology mappings. We introduce USX-PGD (Uncertainty-aware, Sparse, and eXplainable Proper Generalised Decomposition), a non-intrusive ROM for two-phase immiscible flow that addresses all three gaps within a single greedy Alternating Least Squares framework. USX-PGD is benchmarked against Proper Orthogonal Decomposition (POD), standard Proper Generalised Decomposition (PGD), and an intermediate Uncertainty-aware Sparse PGD (US-PGD) variant, on a formation-aware upscaled coarse-grid ($30\times 110\times 34$ cells) representation of the SPE10 Model 2 benchmark, a heterogeneous two-phase reservoir with permeability contrasts spanning six orders of magnitude. US-PGD adds sparsity-promoting thresholding and bootstrap resampling to certify a confidence interval on reconstruction accuracy; USX-PGD further adds formation energy decomposition, mode dominance mapping, breakthrough attribution, and mode sensitivity indexing, attributing the reduced-order modes to identifiable geological formations. All four methods reproduce the reference production curves to within $11.03$–$11.05\%$ NRMS at online reconstruction times of 51–63 ms; the sparse and explainable variants achieve comparable accuracy while additionally providing $41.8\%$ spatial sparsity and a certified $95\%$ confidence interval. All four ROMs compress and replay an already-simulated trajectory, not predict new, unsimulated scenarios; USX-PGD is offered as a reproducible, physically transparent foundation for such multi-query workflows, with predictive extension identified as future work. Full article

Due to its high content of bioactive constituents and associated health benefits, Aronia melanocarpa is considered a superfood, and its consumption has increased substantially in recent years. This growing demand has led to the generation of large quantities of processed by-products, which remain rich in valuable phytochemicals and require sustainable utilization. In this study, four extraction techniques—conventional solvent extraction (CSE), pressurized liquid extraction (PLE), pulsed electric field extraction (PEF), and ultrasound-assisted extraction (UAE)—were comparatively evaluated and optimized for the recovery of bioactive compounds from aronia pomace. A second-order polynomial model (Fit Least Squares) was applied to determine the optimal conditions for each technique. The optimized extracts exhibited distinct phytochemical profiles: total polyphenol content reached 71.5 (UAE), 70.0 (CSE), 67.0 (PLE), and 48.0 (PEF) mg GAE/g dw, while total anthocyanins were 15.5 (UAE), 13.0 (CSE), 11.0 (PEF), and 3.5 (PLE) mg CyE/g dw. Antioxidant capacity ranged from 538.0 to 800.0 µmol AAE/g dw (FRAP) and 17.0 to 39.0 mmol AAE/g dw (DPPH). HPLC analysis confirmed cyanidin-3-O-glucoside as the predominant compound, with concentrations of 5.0 (UAE), 4.6 (CSE), 3.2 (PEF), and 0.8 mg/g dw (PLE). Overall, ultrasound-assisted extraction under optimal conditions provided the highest recovery of bioactive constituents, highlighting its suitability for the valorization of aronia by-products. Full article

This study investigates how quality tourism supply contributes to destination competitiveness through the sequential mediating roles of local resource integration and perceived authenticity within a sustainability-oriented tourism framework. Building upon the Resource-Based View and destination competitiveness theory, the study proposes and empirically tests a sequential mediation model in which local resource integration and perceived authenticity explain the mechanisms through which quality tourism supply contributes to destination competitiveness. Data were collected through a survey of 430 tourism stakeholders, including representatives of hotels, restaurants, tourism boards, travel agencies, rural tourism enterprises, and local producers. The data were analyzed using Partial Least Squares Structural Equation Modeling (PLS-SEM) implemented in SmartPLS 4. The findings indicate that quality tourism supply positively influences local resource integration and destination competitiveness, while local resource integration significantly enhances perceived authenticity. Perceived authenticity, in turn, positively affects destination competitiveness. The results also confirm significant individual and sequential mediation effects, demonstrating that tourism quality generates stronger competitive advantages when destinations effectively integrate local resources and foster authentic tourism experiences. The proposed model demonstrated substantial explanatory power, explaining 45.9% of the variance in local resource integration (R² = 0.459), 50.7% of the variance in perceived authenticity (R² = 0.507), and 63.1% of the variance in destination competitiveness (R² = 0.631). The study contributes to the destination competitiveness literature by introducing local resource integration as a strategic mediating construct and by integrating quality, authenticity, and competitiveness within a unified analytical framework. The findings provide theoretical contributions to destination competitiveness research and practical implications for destination managers and policymakers seeking to promote sustainable and competitive tourism development through place-based and community-oriented approaches. Full article

Underutilised indigenous crops (UICs) are increasingly recognised for their potential to contribute to climate-resilient agriculture, dietary diversification, and household food security among smallholder farmers in South Africa. However, evidence regarding farmers’ knowledge, attitudes, and practices (KAP) toward these crops remains limited. This study examined the KAP of smallholder farmers toward UICs and assessed factors associated with their adoption in selected municipalities of South Africa. A quantitative cross-sectional survey was conducted among 205 smallholder farming households using a structured questionnaire. Respondents were selected through a multi-stage sampling procedure involving purposive, snowball, and stratified sampling. Data were analysed using descriptive statistics, Ordinary Least Squares (OLS) regression, binary logistic regression, and bootstrap mediation analysis. Results showed that smallholder farmers generally possessed positive knowledge and attitudes toward UICs, particularly regarding their nutritional value, adaptability to local climatic conditions, and contribution to household food security. However, utilisation practices remained uneven across households. Age, farming experience, water availability, extension support, and market access were significantly associated with adoption. Formal bootstrap mediation analysis (1000 simulations) did not support attitudinal mediation of the knowledge-adoption relationship (ACME = −0.007, 95% CI: −0.079 to 0.059, p = 0.802), suggesting that knowledge operates through mechanisms other than attitudinal change. The study concludes that UICs possess significant potential to contribute to food security and climate adaptation in the study areas, but stronger institutional support, extension services, and market development are required for wider adoption. Full article

Predicting the pore structure of lignocellulosic biochar from pyrolysis conditions without exhaustive experimental characterization remains an open challenge. We fit semi-empirical Arrhenius-type and power-law correlations linking pyrolysis temperature to $S_{\mathrm{BET}}$, $V_{\mathrm{T}}$, and ${\overline{d}}_p$ by nonlinear least squares, using 45 literature records from seven open access studies (12 feedstocks, 300–800 °C). The central finding is that feedstock category, not temperature alone, dominates variance in $S_{\mathrm{BET}}$: pooled calibration explains only $R^2=0.199$ (RMSE = 183 m² g⁻¹), whereas feedstock-stratified fitting recovers accuracy (e.g., RMSE = 43.6 m² g⁻¹ for grasses, a within-study estimate from a single source). The Arrhenius and power-law forms are statistically indistinguishable ($\Delta \mathrm{AIC}<2$); the Arrhenius form is adopted for physical interpretability. Pooled fits reach $R^2=0.691$ ($V_{\mathrm{T}}$) and $0.563$ (${\overline{d}}_p$). Leave-one-study-out cross-validation (RMSE = 184 m² g⁻¹) confirms that reliable prediction requires calibration data within the target feedstock category. The correlations are descriptive tools valid within their calibration envelope, not general predictive models. Estimation uses no machine learning; a benchmark against OLS and random forest models confirms that greater flexibility improves in-sample fit but not out-of-sample generalization. Full article

Spectroelectrochemistry (SEC) provides unique information-rich datasets by coupling molecular spectroscopic fingerprints with electrochemical activity. Despite its richness, direct machine-learning (ML) analysis of SEC data under realistic experimental constraints and scarcity remains unexplored. This work examines how much spectroscopically encoded electrochemical information can be learned from minimal SEC training data in a chemically reversible two-electron redox system, and how model choice interacts with limited experimental diversity across scan rate. Using purely experimental SEC datasets collected at four scan rates (2, 3, 5, and 7 mV/s), partial least squares (PLS) and convolutional neural networks (CNNs) regressors are evaluated under several SEC-level training, validation, and test configurations. Model performance is assessed across three targets of increasing physical complexity, including species concentrations, derivative cyclic voltabsorptometry (DCVA) current, and experimental cyclic voltammetry (CV) current. Under single-SEC training, both models achieve the expected near-quantitative concentration prediction (R² ~0.99), while performance decreases for DCVA (R² ~0.93) and most substantially for CV current (R² ~0.78), reflecting the progressively weaker and more indirect encoding of these targets within the absorbance data. Introducing minimal experimental diversity with only two distinct training SEC datasets enables both PLS and CNN models to generalize strongly to an unseen third SEC dataset, achieving maximum CV R² values approaching ~0.98 in the most favorable configurations. CNN models extend the apparent linear performance ceiling observed for PLS by capturing localized, scan-rate-conditioned nonlinear correlations between spectral evolution and the experimentally measured CV response, yielding improved waveform reconstruction and greater robustness to training SEC dataset selection. These results demonstrate that, within the present chemically reversible and spectroscopically well-resolved SEC system, high-fidelity prediction of electrochemical targets can be achieved without large datasets when limited but strategically selected electrochemical diversity is introduced. SEC dataset linearity is further shown to be target-dependent and becomes operationally meaningful only when scan-rate space is sufficiently sampled. More broadly, this work establishes a controlled framework for investigating ML-enabled SEC dataset analysis under experimentally scarce conditions and provides guidance for experimental design and calibration in low-data spectroelectrochemical settings. Full article

Algorithmic management is central to platform gig work, but its distinct forms of control may relate differently to worker strain. We examine these relationships in food delivery, a typical form of platform gig work. In this setting, algorithms continuously allocate tasks, track performance, evaluate workers, and shape behavior. The Stressor–Strain–Outcome (S-S-O) model provides the overarching framework. The Job Demands–Resources (JD-R) perspective explains why the three control dimensions may show different associations, while cognitive appraisal theory provides a supplementary lens. Specifically, the study links standardized guidance, tracking evaluation, and behavioral constraints to work engagement through role stress and emotional exhaustion, with overall social support as a moderator. Survey data from 608 food-delivery workers were analyzed using partial least squares structural equation modeling (PLS-SEM) with 5000 bootstrap resamples. Standardized guidance was negatively associated with role stress and emotional exhaustion, whereas tracking evaluation and behavioral constraints were positively associated with both. Role stress and emotional exhaustion each carried significant indirect associations with work engagement; the serial indirect associations were significant but small. Overall social support strengthened the positive associations of tracking evaluation and behavioral constraints with both strain variables. For standardized guidance, however, higher support strengthened, rather than weakened, the negative associations with strain. Predictive assessment, control-variable models, and covariance-based SEM sensitivity analyses produced broadly consistent findings. These results show that distinct forms of algorithmic control relate differently to strain. Because the data are cross-sectional and were collected from food-delivery workers, causal and temporal inferences, as well as generalization across occupations, remain limited. Full article

Accurate identification of customer transformer relationships in distribution sub-zones is a fundamental prerequisite for the refined management of low-voltage distribution networks and the integration of distributed generation sources. Addressing current issues such as missing records, non-standard wiring and unclear boundaries between multiple sub-zones, this paper proposes an identification method based on the Power Admittance Regression Algorithm (PARA). Based on the fundamental laws of electrical circuits, this method constructs a regressible model of the linear relationship between the total admittance at the transformer end and the admittances at each consumer end. By utilising electrical data collected simultaneously from smart metres and distribution transformer terminals, it formulates the identification of consumer transformer relationships as a problem of minimising regression residuals. For three typical operating conditions—pure residential load, mixed residential and commercial load, and photovoltaic connection at the feeder terminus—constrained least-squares regression models and binary regression models incorporating PV variables were established respectively; ridge regression regularisation was introduced to suppress multicollinearity and enhance model robustness. Simulation tests were conducted using a dataset comprising 150 consecutive time sections and 70 test nodes (of which 60 were customers within the local substation area and 10 were interference nodes from other substation areas) for validation. The results indicate that, under the three conditions described above, in engineering simulations accounting for three-phase imbalance, random perturbations in line parameters and measurement noise, the average accuracy of this method, as determined by 100 Monte Carlo simulations, was 86.2 percent, 92.8 percent and 93.1 percent respectively, with standard deviations ranging from 1.6% to 1.9%, thereby validating its effectiveness and superiority in scenarios involving complex load structures and the integration of renewable energy. As this work is based on simulation data, further online validation using actual feeder data from electricity consumption data acquisition systems is required. Full article

This paper presents the mathematical modeling, numerical implementation, and experimental validation of a decoupling control algorithm for a five-port multiport modular active bridge (MMAB) converter in DC microgrid applications. Based on an analytically derived generalized state-space framework of the MMAB topology, a matrix-based method for suppressing non-linear mutual cross-couplings among individual ports is proposed. The study addresses parametric uncertainties within the system matrix caused by parasitic bus inductances; by formulating a linear system of equations solved via the numerical least-squares method, the equivalent parameter identification error was reduced from over 18% to a valid threshold. The decoupling performance and dynamic responsiveness of the closed-loop system were experimentally verified on a dual-core TMS320F28379D digital signal processor. The experimental results demonstrate that the proposed algorithm effectively isolates transient step-load perturbations, maintaining voltage stability on adjacent undisturbed ports within a strict deviation of less than $+0.51\%$ and achieving a recovery time below 5 ms. Furthermore, the real-time execution of the online Jacobian matrix inversion via the Newton–Raphson method confirms the computational feasibility and convergence of the iterative approach under tight sampling periods. The obtained results provide a robust, experimentally validated foundation for advanced algebraic and numerical control strategies in high-stability multiport power conversion systems. Full article

To address the difficulty in accurately characterizing the fault current response of renewable energy grid-connected devices during high-/low-voltage ride-through, this paper proposes a fault current response modeling and parameter identification method based on adaptive nonlinear compensation. First, with the fault voltage and pre-fault operating point as input variables, a basic quadratic equivalent model is established to describe the main variation characteristics of active and reactive currents during high-/low-voltage ride-through. Second, nonlinear compensation terms are introduced into the basic model to correct the response deviation caused by the simplification of fast electromagnetic control links in the electromechanical transient equivalent process, thereby improving the representation capability of the model for complex fault current characteristics. Furthermore, considering that the structural parameters of the nonlinear compensation terms are difficult to directly identify using the traditional least squares method, a differential evolution–ridge regression (DE–Ridge) hierarchical identification method is proposed. In this method, the differential evolution algorithm is used in the outer layer to adaptively optimize the nonlinear structural parameters, while ridge regression is used in the inner layer to solve the corresponding linear coefficients. Case study results show that, compared with the traditional quadratic equivalent model and the fixed nonlinear compensation model, the proposed method further reduces the fault current identification error on the validation set and improves the identification accuracy and generalization capability of fault current responses during high-/low-voltage ride-through. Full article