Search Results (461)

This research investigates the influence of economic, energy, and institutional variables on sustainable economic growth for Pacific Asian countries using Adjusted Net Savings (ANS) as a more refined measure of sustainable development. Using an unbalanced panel dataset for the period 1996 to 2021, second-generation panel data analysis is conducted to capture both long-run and distributional relationships, addressing potential concerns about cross-sectional dependence. The results indicate the presence of long-run relationships that are stable for both sustainable development itself and for its defining factors. Foreign direct investments (FDI) are found to have the most significant influence on sustainable development for all quantile values, underlining their central importance to long-run capital accumulation efforts. Renewable energy consumption helps increase sustainability outcomes for countries with lower savings performance values, while renewable energy production is found to have a modest but positive influence for each quantile of the distribution of outcomes. Natural resource wealth is seen to have non-linear effects on outcomes, with countries with lower savings values being adversely affected, while countries with higher savings values are beneficially affected. The presence of institutional factors is an enabler for countries with lower values of sustainable development performance. Full article

Unbalanced supply voltage (USV) represents a critical power quality challenge in industrial environments, significantly degrading the performance, efficiency, and operational lifespan of three-phase induction motors. Accurate real-time estimation of sequence impedances ($Z_a,Z_b,Z_c$) and detection of the Negative Voltage Factor (NVF) are essential for effective condition monitoring and preventive maintenance strategies. While existing machine learning methods have demonstrated promising accuracy, they often rely on manual feature engineering, lack hierarchical representation learning, and treat impedance estimation and fault detection as isolated tasks. This paper proposes a unified Deep Multi-Task Learning framework that leverages Residual Multilayer Perceptron (ResMLP) architectures for feature-based learning and Temporal Convolutional Networks (TCNs) for end-to-end raw signal learning. Our contributions include: (1) introduction of a Multi-Head ResMLP architecture that jointly optimizes phase impedance and fault detection, achieving superior NVF accuracy (MAE = 0.0007) and a fault detection F1-score of 0.8831; (2) investigation of raw-voltage TCN models for voltage-only diagnostics, with analysis of the trade-offs between end-to-end learning and feature-based approaches; (3) extensive ablation studies demonstrating the impact of network depth, data augmentation, and training protocols on model generalization; and (4) deployment of PyTorch (v2.0.1)-based models suitable for embedded systems with real-time inference capabilities (2.3 ms per prediction). Experimental validation on a 1.1 kW three-phase motor dataset under diverse load conditions (0–10 Nm) and USV magnitudes (5–15 V) confirms the robustness and practical applicability of the proposed approach for industrial fault diagnosis and condition monitoring systems. Full article

Image restoration has wide applications in the field of computer vision, yet existing methods suffer from limitations. CNNs struggle to capture long-range dependencies, while transformers exhibit insufficient performance in handling local details and high computational complexity. Additionally, existing dual-branch networks fail to define a clear dominant–auxiliary role between branches, leading to redundancy and high computational costs. This paper proposes a Global Modulated Unbalanced Dual-Branch Network (GMUD-Net), which innovatively adopts an unbalanced structure with a CNN as the main branch and a transformer as the auxiliary branch. Specifically, the CNN branch achieves strong restoration capability by integrating the global–local hybrid backbone block (GLBB) and the frequency-based global attention module (FGAM). As the key building block in the CNN branch, GLBB integrates a local backbone branch, a global Fourier branch, and a residual branch to fuse local details with global context. Meanwhile, FGAM leverages the fast Fourier transform at the bottleneck to enhance cross-channel interaction and improve global restoration performance. In addition, the lightweight transformer branch employs efficient cross-channel attention to provide complementary global cues, which are filtered and injected into the CNN branch via the global attention guidance block (GAG). These designs integrate the advantages of both CNNs and transformers while significantly reducing computational burden, offering a new paradigm to address the limitations of traditional dual-branch architectures. Experimental results demonstrate that compared with existing algorithms, the proposed method achieves state-of-the-art or highly competitive performance in both quantitative evaluations and qualitative results across nine datasets. Full article

Predictive pest management in urban sewer networks represents a sustainable alternative to reactive, biocide-based methods. Using data collected through an IoT architecture and validated with manual inspections across eight manholes over 113 days, we implemented a rigorous comparative framework evaluating eleven data mining algorithms, including classical methods (KNN, SVM, decision trees) and advanced ensemble techniques (XGBoost, LightGBM, CatBoost) optimized for unbalanced datasets. Gradient boosting models with explicit handling of class imbalance—where the absence of pests exceeds 77% of observations—showed exceptional performance, achieving a Macro-F1 score above 0.92 and high precision in identifying the minority high-risk class. Explainability analysis using SHAP consistently revealed that elevated CO₂ concentrations are the primary predictor of infestation, enabling early identification of critical zones. This study demonstrates that carbon dioxide (CO₂) acts as the most robust bioindicator for predicting severe infestations of Periplaneta americana, significantly outperforming conventional environmental variables such as temperature and humidity. The implementation of the model in a real-time monitoring platform generates interpretable heat maps that support proactive and localized interventions, optimizing resource use and reducing dependence on biocides. This study presents a scalable, operationally viable predictive system designed for direct integration into municipal asset management workflows, offering a concrete, industry-ready solution to transform pest control from a reactive, labor-intensive process into a data-driven, proactive operational paradigm. This approach not only transforms pest management from reactive to predictive but also aligns with the Sustainable Development Goals, offering a scalable, interpretable, and operationally viable system for smart cities. Full article

Sustainable corporate governance plays an essential role in promoting responsible economic growth and enhancing social and environmental well-being in emerging economies. In this context, Environmental, Social, and Governance (ESG) performance has become an important indicator of a firm’s commitment to sustainable development and its alignment with the United Nations Sustainable Development Goals, particularly SDG 8 and SDG 12. This study investigates the impact of Environmental, Social, and Governance (ESG) performance on the financial sustainability of publicly listed companies in Thailand, a rapidly developing Southeast Asian economy where empirical evidence remains limited. Using an unbalanced panel dataset of 965 firm-year observations across multiple industries, multiple regression models were employed to assess the influence of ESG performance on two financial indicators: return on capital employed and return on assets. Granger causality tests were also conducted to explore directional relationships between sustainability performance and financial outcomes. The empirical results reveal a significant negative short-term association between ESG performance and return on assets (ROA), whereas the relationship with return on capital employed (ROCE) is statistically insignificant. The causality analysis indicates that ESG performance Granger-causes ROA, implying that sustainability-driven strategic decisions may precede and influence financial outcomes over time. Additionally, leverage emerges as a key constraint to financial sustainability, negatively affecting both ROCE and ROA. These findings underscore the challenge of striking a balance between sustainability investments and immediate profitability in emerging markets. Policymakers and business leaders are encouraged to promote supportive governance frameworks, reduce financial barriers, and foster ESG-driven practices that contribute to long-term sustainable competitiveness and inclusive development. Full article

This study empirically examines how artificial intelligence (AI) adoption and innovation shape product market dynamics and financial performance in multinational enterprises (MNEs) using a global firm sample over 1980–2023. We construct an unbalanced panel dataset by integrating textual analysis, manual verification, and data merged from nine major databases, identifying 411 AI-classified MNEs and a matched 411 non-AI MNEs. Using panel regression models with industry and year fixed effects, we test how AI intensity (the proportion of AI-related products/assets) and R&D—individually and jointly—affect product portfolio breadth and change, market share, industry concentration (HHI), and profitability. The results show that greater AI integration is associated with higher product diversification and a stronger competitive positioning, and that the interaction of AI and R&D is particularly important for explaining market share, concentration, and profitability differences across AI and non-AI MNEs. Overall, the findings highlight the strategic value of aligning AI adoption with innovation investments to strengthen product market outcomes and financial performance in global markets. Full article

Background/Objectives: Fetal health is essential in prenatal care, influencing both maternal and fetal outcomes. Cardiotocography (CTG) monitors uterine contractions and fetal heart rate, yet manual interpretation exhibits significant inter-examiner variability. Machine learning offers automated alternatives; however, class imbalance in CTG datasets where pathological cases constitute less than 10% leads to poor detection of minority classes. This study aims to provide the first systematic benchmark comparing five resampling strategies across seven classifier families for multi-class CTG classification, evaluated using imbalance-aware metrics rather than overall accuracy alone. Methods: Seven machine learning models were employed: Naïve Bayes (NB), Random Forest (RF), Linear Discriminant Analysis (LDA), k-Nearest Neighbors (KNN), Linear Support Vector Machine (SVM), Multinomial Logistic Regression (MLR), and Multi-Layer Perceptron (MLP). To address class imbalance, we evaluated the original unbalanced dataset (base) and five resampling methods: SMOTE, BSMOTE, ADASYN, NearMiss, and SCUT. Performance was evaluated on a held-out test set using Balanced Accuracy (BACC), Macro-F1, the Macro-Matthews Correlation Coefficient (Macro-MCC), and Macro-Averaged ROC-AUC. We also report per-class ROC curves. Results: Among all models, RF proved most reliable. Training on the original distribution (base) yielded the highest BACC (0.9118), whereas RF combined with BSMOTE provided the strongest class-balanced performance (Macro-MCC = 0.8533, Macro-F1 = 0.9073) with a near-perfect ROC-AUC (approximately 0.986–0.989). Overall, resampling effects proved model dependent. While some classifiers achieved optimal performance on the natural class distribution, oversampling techniques, particularly SMOTE and BSMOTE, demonstrated significant improvements in minority class discrimination and class-balanced metrics across multiple model families. Notably, certain models benefited substantially from resampling, exhibiting enhanced Macro-F1, BACC, and minority class recall without sacrificing overall accuracy. Conclusions: These findings establish robust, model-agnostic baselines for CTG-based fetal health screening. They highlight that strategic oversampling can translate improved minority class discrimination into clinically meaningful performance gains, supporting deployment in cost-sensitive and threshold-aware clinical settings. Full article

This research examines the influence of environmental, social, and governance (ESG) strategies on corporate financial performance (CFP) in China’s automotive industry, characterized by intense regulatory pressure and fast-paced technological transformation. Using an unbalanced panel dataset of A-share listed automotive firms from 2009 to 2024, this paper combines ESG scores from the Huazheng ESG index with firm-level financial data from CSMAR. CFP is measured through both accounting-based (ROA) and market-based (Tobin’s Q) indicators. Panel regression models are applied to evaluate the influence of overall ESG performance and the three individual pillars, and to assess heterogeneity across ownership types, firm type, and firm age. The results show that ESG performance is significantly and positively associated with ROA, but is insignificantly associated with Tobin’s Q. It is suggested that ESG engagement improves accounting profitability but is not fully reflected in the capital market. Among the three ESG pillars, governance shows the strongest positive link with ROA, while environmental and social performance are weakly associated with ROA. Furthermore, the heterogeneity study shows that the positive relationship between ESG and CFP is more pronounced for non-state-owned firms, vehicle manufacturers, or mature firms. Overall, this paper presents fresh evidence on whether and how ESG initiatives can facilitate sustainable value in China’s automotive sector, offering insights for policymakers and management that may help this industry achieve sustainable growth. Full article

Data classification tasks based on deep neural networks and machine learning are increasingly used in different fields, such as medicine, finance, and data circulation. However, in these applications, the accuracy of predictions must be guaranteed, and the privacy and security of prediction data and models must be guaranteed. In an unsafe cloud environment, cloud users are reluctant to use the classification prediction tasks provided by the cloud. To solve these problems, this paper researches the data oversampling method and proposes the G-MSMOTE method, which can solve the oversampling problem of multiple minority classes in the data set, generate more diverse data, and solve the data imbalance problem. By improving the traditional FV and using CRT technology to improve coding efficiency, the cloud receives the user’s encrypted ciphertext, and the neural network completes the data prediction task in the ciphertext, thereby providing confidentiality for user data and model parameters under the semi-honest adversarial model, assuming the security of the underlying fully homomorphic encryption scheme and accepting the leakage of model architecture and ciphertext sizes. The feasibility of our method was demonstrated through experimental comparative analysis. We created unbalanced cases based on the MNIST dataset and performed comparative analysis in plain and ciphertext. In the balanced dataset, the model’s prediction accuracy in ciphertext reached 93.44%. In the unbalanced case, after preprocessing with our improved G-MSMOTE algorithm, the model’s prediction accuracy in ciphertext increased by at least 10%. These results show that our scheme can efficiently, accurately, and securely (under the semi-honest model) complete the data classification prediction task. Full article

Reconstructing original cusp dimensions in worn molars represents a fundamental challenge across dentistry, anthropology, and paleontology, as dental wear obscures critical morphological information. In this proof-of-concept study, we present a standardized machine learning pipeline for predicting original cusp height, specifically the horn tips of the enamel–dentine junction (EDJ), in worn lower molars using three-dimensional morphometric data from micro-computed tomography (micro-CT). We analyzed 40 permanent lower first (M1) and second (M2) molars from four hominin groups, systematically evaluated across three wear stages: original, moderately worn (worn1), and severely worn (worn2). Morphometric variables including height, area, and volume were quantified for each cusp, with Random Forest and multiple linear regression models developed individually and combined through ensemble methods. To mimic realistic reconstruction scenarios while preserving a known ground truth, models were trained on unworn specimens (original EDJ morphology) and tested on other teeth after digitally simulated wear (worn1 and worn2). Predictive performance was evaluated using root mean square error (RMSE) and coefficient of determination (R²). Our results demonstrate that under moderate wear (worn1), the ensemble models achieved normalized RMSE values between 11% and 17%. Absolute errors typically below 0.25 mm for most cusps, with R² values up to ~0.69. Performance deteriorated under severe wear (worn2), particularly for morphologically variable cusps such as the hypoconid and entoconid, but generally remained within sub-millimetric error ranges for several structures. Random Forests and linear models showed complementary strengths, and the ensemble generally offered the most stable performance across cusps and wear states. To enhance transparency and accessibility, we provide a comprehensive, user-friendly software pipeline including pre-trained models, automated prediction scripts, standardized data templates, and detailed documentation. This implementation allows researchers without advanced machine learning expertise to explore EDJ-based reconstruction from standard morphometric measurements in new datasets, while explicitly acknowledging the limitations imposed by our modest and taxonomically unbalanced sample. More broadly, the framework represents an initial step toward predicting complete crown morphology, including enamel thickness, in worn or damaged teeth. As such, it offers a validated methodological foundation for future developments in cusp and crown reconstruction in both clinical and evolutionary dental research. Full article

Today, machine learning (ML) is generally considered a potent and efficient tool for addressing studies in various diverse domains, including image processing and event prediction on a timescale. ML represents complex relations between features, and these mappings between such features may be applied in simulations of time-dependent events, such as the behavior of animals. Still, ML inherently strongly depends on extensive and consistent datasets, a fact that reveals both the benefits and drawbacks of ML. In the use of ML, insufficient or skewed data can limit the ability of algorithms to accurately predict or generalize possible states. To overcome this limitation, this work proposes an integrated hybrid approach that combines machine learning with methods from cognitive science, here especially inspired by the ACT-R model to approach cases of missing or unbalanced data. By incorporating cognitive processes such as memory, perception, and attention, the model accounts for the internal mechanisms of decision-making and environmental interaction where traditional ML methods fall short. This approach is particularly useful in representing states that are not directly observable or are underrepresented in the data, such as rare behavioral responses for animals, or adaptive strategies. Experimental results show that the combination of machine learning for data-driven analysis and cognitive ‘rule-based’ frameworks for filling in gaps provides a more comprehensive model of animal behavior. The findings suggest that this hybrid approach to simulation models can offer a more robust and consistent way to study complex, real-world phenomena, especially when data is inherently incomplete or unbalanced. Full article

Understanding patterns and mechanisms of species diversity is one fundamental issue in biogeography and ecology. As a critical region for biodiversity, the Qinghai-Xizang Plateau (QXP) still has unclear distribution patterns and drivers for cryptic, understudied taxa such as Curculionoidea. Here, we collected the distribution data of Curculionoidea on the QXP to analyze their diversity patterns and influencing factors, and compiled a DNA barcode dataset to uncover cryptic diversity. This comprehensive dataset encompasses 671 Curculionoidea species across 223 genera, demonstrating a level of diversity that surpasses that of certain vertebrate groups. We also observed an unbalanced biogeographic pattern of diversity, with a concentration of species in the eastern and southern regions and a scarcity in the northern and central areas of QXP. Further analysis showed that the elevation range is the most important factor influencing the diversity of Curculionoidea. In addition, based on 1147 COI-5′ barcode sequences from 217 species, we found that 11 morphological species may contain cryptic species based on DNA barcode datadset. Our findings significantly enhance the current understanding of cryptic biodiversity patterns among understudied taxa in the QXP, while simultaneously highlighting persistent knowledge gaps in characterizing the plateau’s full ecological complexity. Full article

This study examines the relationship between national open educational resource (OER) policies and Sustainable Development Goal 4 (SDG4) outcomes across 187 countries between 2015 and 2024, with particular attention to the moderating role of artificial intelligence (AI) readiness. Despite widespread optimism about digital technologies as catalysts for universal education, systematic evidence linking formal OER policy frameworks to measurable improvements in educational access and completion remains limited. The analysis employs fixed effects and difference-in-differences estimation strategies using an unbalanced panel dataset comprising 435 country-year observations. The research investigates how OER policies associate with primary completion rates and out-of-school rates while testing whether these relationships depend on countries’ technological and institutional capacity for advanced technology deployment. The findings reveal that AI readiness demonstrates consistent positive associations with educational outcomes, with a ten-point increase in the readiness index corresponding to approximately 0.46 percentage point improvements in primary completion rates and 0.31 percentage point reductions in out-of-school rates across fixed effects specifications. The difference-in-differences analysis indicates that OER-adopting countries experienced completion rate increases averaging 0.52 percentage points relative to non-adopting countries in the post-2020 period, though this estimate remains statistically imprecise (p equals 0.440), preventing definitive causal conclusions. Interaction effects between policies and readiness yield consistently positive coefficients across specifications, but these associations similarly fail to achieve conventional significance thresholds given sample size constraints and limited within-country variation. While the directional patterns align with theoretical expectations that policy effectiveness depends on digital capacity, the evidence should be characterized as suggestive rather than conclusive. These findings represent preliminary assessment of policies in early implementation stages. Most frameworks were adopted between 2019 and 2022, providing observation windows of two to five years before data collection ended in 2024. This timeline proves insufficient for educational system transformations to fully materialize in aggregate indicators, as primary education cycles span six to eight years and implementation processes operate gradually through sequential stages of content development, teacher training, and institutional adaptation. The analysis captures policy impacts during formation rather than at equilibrium, establishing baseline patterns that require extended longitudinal observation for definitive evaluation. High-income countries demonstrate interaction coefficients between policies and readiness that approach marginal statistical significance (p less than 0.10), while low-income subsamples show coefficients near zero with wide confidence intervals. These patterns suggest that OER frameworks function as complementary interventions whose effectiveness depends critically on enabling infrastructure including digital connectivity, governance quality, technical workforce capacity, and innovation ecosystems. The results carry important implications for how countries sequence educational technology reforms and how international development organizations design technical assistance programs. The evidence cautions against uniform policy recommendations across diverse contexts, indicating that countries at different stages of digital development require fundamentally different strategies that coordinate policy adoption with foundational capacity building. However, the modest short-term effects and statistical imprecision observed here should not be interpreted as evidence of policy ineffectiveness, but rather as confirmation that immediate transformation is unlikely given implementation complexities and temporal constraints. The study contributes systematic cross-national evidence on aggregate policy associations while highlighting the conditional nature of educational technology effectiveness and establishing the need for continued longitudinal research as policies mature beyond the early implementation phase captured in this analysis. Full article

Environmental regulations serve as a critical determinant of industrial competitiveness in the global market. Recent policy shifts have driven a gradual convergence of rural environmental standards with urban norms, fostering a dynamic landscape of “top-down competition” between urban and rural regulatory frameworks. While the economic consequences of regional regulatory disparities are well-documented, the specific impacts of this regulatory convergence remain insufficiently explored. To address this gap, this study constructs a novel index to measure the convergence of environmental regulations between urban districts and rural counties at the prefecture level. Utilizing an unbalanced panel dataset of 5600 county-level timber processing enterprises, the Heckman two-stage model is employed for empirical analysis. The results demonstrate that the convergence of urban and rural environmental regulations significantly enhances both the export probability and export intensity of county-level firms, with these effects exhibiting persistence and cumulative growth over time. These findings remain robust across a series of validation tests, including instrumental variable estimation, double machine learning, and alternative model specifications. Mechanism analysis reveals that regulatory convergence promotes exports primarily by improving access to green credit and enhancing peer quality within the industry. Furthermore, heterogeneity tests indicate that the positive effects are most pronounced for start-ups and firms in the decline stage, as well as for enterprises located in eastern China, those outside the Yangtze River Economic Belt, and those subject to minimal government intervention. This study provides critical micro-level evidence that helps enterprises navigate the evolving policy landscape and supports the formulation of strategies to boost export trade amidst the integration of environmental regulations. Full article

Institutional quality and globalization are crucial in influencing both the prevalence and quality of sustainable entrepreneurial ecosystems within an economy. This study examines the relationship between Opportunity-Driven Entrepreneurship (ODE); entrepreneurial quality, as measured by the Motivational Index (MI), and institutional quality, assessed through economic freedom and governance, in high- and middle-income countries. It also examines how globalization impacts both ODE and MI in these country groups. Using data from the Global Entrepreneurship Monitor (GEM) and combined indices of economic freedom, governance, and globalization, the study analyzes an unbalanced panel dataset comprising 64 countries from 2004 to 2018. Estimation is performed using the Robust Least Squares (RLS) method. The findings show that economic freedom has a positive and significant effect on both ODE and MI across high- and middle-income countries. In contrast, governance has a significant impact on ODE and MI only in high-income countries. Globalization exerts a negative influence on ODE across both income groups, with the adverse effect being more pronounced in middle-income countries. Conversely, its effect on MI is positive in middle-income countries but shows no significant influence in high-income economies. The study offers valuable insights for economists, policymakers, and scholars interested in the forces that shape ODE. Full article