Search Results (274)

Benchmarking deep learning forecasters against classical and physically based numerical baselines remains uncommon in the time-series forecasting literature. Meteorological station networks offer an under-exploited evaluation environment, uniquely providing a physically based climate-model comparator alongside standard baselines. We evaluated eight forecasting approaches—climatology, SARIMA, Random Forest, and five deep learning architectures (TFT, N-HiTS, PatchTST, TiDE, xLSTM)—against bias-corrected output from a five-member CMIP6 ensemble, on 100 meteorological stations across four Western Balkan countries (monthly temperature and precipitation, 1961–2020), using non-parametric significance testing, a rolling-origin backtest (five windows, 2011–2020), and a five-seed robustness check. For temperature, all five deep learning architectures achieved lower MAE than the classical baselines (p < 10⁻⁹⁹), though PatchTST’s advantage over climatology was not significant; the best-performing architecture varied across seeds and evaluation windows, so we characterise a leading cluster (N-HiTS, TFT, TiDE, PatchTST) rather than a single winner. The primary temperature advantage was geographically broad-based, while the comparison against the physical-model baseline was robust to the choice of comparator GCM. For precipitation, by contrast, a simple climatological-mean baseline outperformed all five deep learning architectures with no exception across all five rolling-origin windows. The deep learning advantage over classical and physical baselines is thus variable-specific rather than universal. Meteorological station networks, combined with a physically based climate-model comparator, constitute a well-suited evaluation environment for the broader time series forecasting community. Full article

With the widespread adoption of cloud computing technology, modern cloud platforms have become increasingly complex and dynamic, posing significant challenges for efficient resource management. Accurate forecasting of cloud resource loads has therefore become essential for improving service quality, optimizing resource utilization, and reducing operational costs. To address the intrinsic characteristics of cloud load time series, including nonlinear fluctuations, multi-scale temporal dependencies, and redundant high-dimensional features, this paper proposes the MST-iTransformer model, which integrates multi-scale temporal encoding, sparse attention, and adaptive feature selection mechanisms. Specifically, a multi-scale temporal encoding module is developed to capture and fuse temporal dependencies across multiple periodic scales. Furthermore, an adaptive feature selection module is introduced to dynamically assign importance weights to resource features, enhancing informative variables while suppressing redundant ones. Meanwhile, a sparse attention mechanism is incorporated to reduce computational overhead while maintaining forecasting accuracy. The proposed model is evaluated on the Alibaba Cluster Trace dataset. Experimental results demonstrate that MST-iTransformer achieves MSE, RMSE, and MAE values of 0.5559, 0.7456, and 0.4968, respectively. Compared with the original iTransformer, the proposed model achieves simultaneous reductions in prediction errors and inference latency, validating the effectiveness of the multi-scale temporal encoding, sparse attention mechanism, and adaptive feature selection modules in improving forecasting accuracy and computational efficiency. These improvements provide reliable prediction support for resource scheduling and elastic scaling in cloud data centers. Full article

This paper investigates the multifaceted macroeconomic drivers of vehicle electrification in India and correlates them with micro-level corporate financial returns using a rigorous dual-stage econometric framework. Stage 1 employs a Newey–West time-series estimator on monthly observations to evaluate aggregate consumer demand elasticities across the automotive sector. Stage 2 utilizes a fixed effects panel specification with clustered standard errors across 10 major Indian automotive manufacturers over a four-year fiscal horizon. Stage 1 results demonstrate that short-run variations in Brent crude prices lack joint predictive power over domestic retail metrics ($F=0.89,p=0.4166$), supporting the thesis that state-owned OMC price-smoothing insulates short-term market dynamics from global oil shocks. Conversely, Stage 2 panel estimations prove that annual global Brent crude fluctuations yield no significant contemporaneous margin shocks. However, expanding annual EV market penetration exerts a substantive negative impact ($\beta =-2.49,p=0.107$) bordering statistical significance on corporate operating profit margins. This operational decoupling reflects a prominent industry ‘J-curve’, where accelerating consumer adoption cycles are countered by heavy front-loaded capital expenditures, asset re-tooling, and unoptimized economies of scale. These findings provide critical direct and indirect strategic insights for organizational stakeholders and policymakers navigating transitional capital cycles in emerging markets. Full article

Multiphase flow metering is widely employed in the oil and gas industry, particularly for measuring gas-liquid-solid multiphase flow at drilling outlets. Low-energy gamma flowmeters offer relatively high metering accuracy, with phase fraction errors for gas, liquid, and solid typically within ±10%. However, in practical applications, fluid viscosity often causes substances to adhere to the photon detector, leading to measurement deviations that can reach 18% or more. To overcome this limitation, this paper proposes a hybrid Density-Based Spatial Clustering of Applications with Noise (DBSCAN)-Transformer regression framework, referred to as D-Transformer. DBSCAN removes isolated abnormal detector responses before overlapping time-series windows are generated, while the Transformer captures temporal dependencies among operating variables, raw phase-fraction readings, and multi-energy photon counts. Under experiment-wise five-fold evaluation, D-Transformer obtains $R^2$ values of 0.982, 0.985, and 0.981 and RMSE values of 0.0134, 0.0122, and 0.0138 for the gas, liquid, and solid phase fractions, respectively. Component ablations and baseline comparisons show that the complete framework outperforms the no-ResNet, no-DBSCAN, CNN-GRU-Attention, CNN-LSTM, ridge-regression, and uncorrected-flowmeter alternatives. Full article

Quantifying the large-scale impact of combined ozone (O₃) and nitrogen dioxide (NO₂) pollution on terrestrial carbon sinks remains a major challenge. Here, we develop a parsimonious yet robust empirical framework that leverages high-resolution remote sensing datasets (CHAP O₃/NO₂ and MODIS NPP, 2008–2021) to characterize nonlinear threshold responses of terrestrial net primary productivity (NPP) across China’s diverse ecosystems. Our observational analysis identifies only associative temporal relationships between annual NPP variability and pollutant concentrations, with NPP positively correlated with O₃ (Pearson’s r = 0.714, p < 0.01) and negatively correlated with NO₂ (r = −0.599, p < 0.05). Notably, the ecosystem-specific threshold values (O₃: 28,324–34,391 μg m⁻³ yr⁻¹; NO₂: 3646–4968 μg m⁻³ yr⁻¹) are statistically derived from spatially aggregated pixel-level records across the full 14-year period, independent of the national annual time-series correlation analyses. Distinct from previous single-pollutant national evaluations, our study advances a novel analytical framework focusing on the interactive and combined impacts of O₃ and NO₂ co-exposure. The results demonstrate that NPP displays an increasing trend under low-level pollutant exposure but declines substantially once pollutant loads exceed the identified threshold ranges. Based on K-means clustering and segmented regression analyses, we estimate a national average NPP reduction of 17.4% per year (−0.68 Pg C yr⁻¹), resulting in a cumulative carbon loss of −9.48 Pg C over the 14-year study period—equivalent to 2.45 years of China’s total terrestrial carbon uptake. Among all ecosystem types, forestlands experience the largest cumulative carbon loss (−4.22 Pg C), with prominent loss hotspots concentrated on the Tibetan Plateau and Northwest China. This refined national-scale assessment of dual-pollutant impacts provides observation-based evidence of substantial terrestrial carbon sink degradation, underscoring the necessity of combined air pollution mitigation strategies to sustain ecosystem stability and climate mitigation targets. Full article

Background and Clinical Significance: Parkinson’s disease (PD) is a progressive neurodegenerative disorder for which nonpharmacological neuromodulatory approaches are being explored as adjunctive strategies. Deep transcranial magnetic stimulation (dTMS) using an H-coil can engage broader and deeper cortical networks than conventional focal coils, although anatomically distant structures are expected to be influenced through network modulation rather than direct electromagnetic stimulation. Case Presentation: We describe a five-patient exploratory case series evaluating clinical measures and cerebral perfusion before and three months after high-frequency dTMS targeting the supplementary motor area (SMA). Clinical outcomes included the Unified Parkinson’s Disease Rating Scale (UPDRS) Parts I-IV, Non-Motor Symptoms Scale (NMSS), Hoehn–Yahr stage, and Timed Up and Go test. Antiparkinsonian medication regimens and doses remained unchanged during the three-month follow-up. Cerebral perfusion was assessed using single-photon emission computed tomography (SPECT) and statistical parametric mapping. No clinical outcome reached nominal statistical significance at follow-up; given the sample size, this should be interpreted as limited statistical power rather than evidence of no effect. At an exploratory voxel-wise threshold of p < 0.001, uncorrected, SPECT identified a 28-voxel cluster of increased regional cerebral blood flow in the right cerebellar cortex (peak MNI coordinates: 22, −38, −44; t = 5.96; z = 3.11). Conclusions: These observations are hypothesis-generating and may reflect network-level modulation of SMA–cerebellar circuitry. Because only five patients were included, no sham-controlled arm was available, and the imaging analysis used an uncorrected exploratory threshold, causal or efficacy claims cannot be made. Larger sham-controlled studies with standardized dose-to-imaging timing, formal neuropsychological assessment, individualized targeting, and prespecified corrected imaging analyses are warranted. Full article

Task-evoked EEG can reveal individual differences in cognitive processing, but it also creates a machine-learning challenge: many epochs are recorded from relatively few participants, and evaluation can be misleading if within-subject dependence is ignored. This study compared handcrafted, ROCKET, and EEGNet representations under subject-wise validation for classifying questionnaire-derived Dark Triad profiles from task-onset EEG. Dark Triad traits were assessed with the Dirty Dozen and clustered into four exploratory multivariate profiles using k-means. EEG was recorded during a visual speeded decision task, and epochs were extracted from −200 to 1000 ms around task onset. The final dataset included 1780 retained task-onset epochs from 30 participants. Three representations were evaluated under identical five-fold subject-wise cross-validation: XGBoost with handcrafted EEG features, XGBoost with ROCKET-derived time-series features, and compact EEGNet. All performance estimates were based only on predictions from held-out participants. The handcrafted model achieved balanced accuracy of 60.9%, whereas ROCKET and EEGNet improved performance to 74.3% and 76.2%, respectively, with only a modest difference between the waveform-based representations. A participant-level label-shuffling analysis of the out-of-fold predictions indicated that prediction–label alignment exceeded chance for all models. Across models, the mean probability assigned to the true cluster increased with psychometric cluster centrality, suggesting that borderline profiles were harder to classify. Signed channel-wise ablation of EEGNet suggested distributed model sensitivity, with the largest positive effects over posterior/parietal electrodes. The findings highlight the importance of participant-level validation, EEG signal representation, and psychometric label structure, while emphasizing the need for external validation in larger independent cohorts. Full article

Compound flooding increasingly threatens developed coastal deltas. High risk pressure does not necessarily produce a resilience deficit where capacity is sufficient, whereas moderate-pressure areas may remain vulnerable when capacity is weak. Recent assessments increasingly integrate hazard, exposure, vulnerability, and adaptive capacity within unified risk frameworks such as the IPCC AR5 risk model. However, by collapsing these dimensions into a single composite risk score, such formulations cannot explicitly diagnose where—and by how much—compound flood risk pressure exceeds intrinsic resilience capacity, which is the information most directly needed for prioritizing resilience investment. This study diagnoses compound flood resilience deficits across Jiangsu Province, China, at county scale from 2000 to 2020. We introduce a diagnostic approach that separates risk pressure from intrinsic resilience capacity and quantifies their spatial mismatch. The risk pressure index is evaluated for consistency with observed disaster-loss indicators—direct economic loss and flood-affected area—over 2010–2020, and spatial statistics, time-series clustering, and explainable machine learning identify deficit patterns, pathways, and associated factors. Both the risk-pressure index and the derived deficit index are positively associated with observed losses, confirming that the framework captures major flood impacts. The resilience deficit index reveals persistent risk–resilience mismatch across Jiangsu. Three pathways emerge: capacity-buffered exposure, inland adaptive adjustment, and coastal resilience-deficit lock-in. Land-system conditions, communication access, transport connectivity, and economic recovery capacity are jointly associated with resilience deficits. The framework offers a transferable approach for prioritizing differentiated flood-risk management in coastal deltas. Full article

Accurate binarization of phase-detection probe signals (gas vs. liquid) is necessary for the estimation of local void fraction, interfacial velocity, and bubble statistics in gas–liquid flows, particularly in nuclear thermal–hydraulic experiments. Classical threshold-based methods—single or double level—perform well on clean laboratory signals but degrade under realistic industrial conditions where noise, baseline drift, and clustered (slug-like) events challenge fixed rules. This work investigates whether deep learning (DL) models trained exclusively on synthetic data can deliver robust, generalizable binarization on real probe measurements. We (i) build a parametric generator of realistic time series from bubbly pulse templates, extended to clusters/slug patterns and perturbed with controlled noise, drift, and oscillatory baselines; (ii) train four lightweight DL architectures—one-dimensional U-Net (UNET-1D), Temporal Convolutional Network (TCN), a minimal one-dimensional Convolutional Neural Network (CNN-1D), and a Bidirectional Long-Short Memory network (BiLSTM)—only on synthetic signals; and (iii) evaluate them against classical threshold methods using event-level and sample-level metrics. On synthetic signal evaluation, UNET-1D and TCN achieve near-perfect event detection and sub-millisecond onset errors. On real bubbly and slug flow sensor data, classical threshold-based methods remain highly competitive on clean sensor signals, while DL models retain advantages under non-stationary baselines and clustered events, yielding accurate void and timing with no hand-tuned assumptions. Results support DL as a practical, data-driven complement to fixed algorithms, particularly in noisy or drift-dominated measuring conditions typical of nuclear thermal–hydraulic loops and safety-relevant test facilities. Full article

This paper conducts a comprehensive empirical evaluation of macroeconomic outcomes under two competing policy paradigms that have governed the United States since the end of World War II: the Keynesian-inflected New Deal Order (1946–1980), rooted in John Maynard Keynes’s theory of aggregate demand management, and the Neoliberal Order (1981–2024), shaped by the price mechanism epistemics of Friedrich A. Hayek and the monetarist counter-revolution of Milton Friedman. Drawing on the historiographical frameworks we operationalize the 1981 policy transition as a natural experiment and apply a fourteen-component analytical pipeline to an annual panel of fifteen Federal Reserve Economic Data (FRED) series spanning 1946–2024. Our methods include: Pruned Exact Linear Time (PELT) structural break detection; normality-adaptive hypothesis testing with Cohen’s d effect sizes; Principal Component Analysis (PCA); a gradient-boosted XGBoost classifier with SHAP interpretability; static and rolling Phillips curve estimation; bootstrapped fiscal multiplier analysis using the corrected Federal Surplus/Deficit ratio; Composite Macroeconomic Performance Index (CMPI) construction; Pearson correlation structure comparison; regime-stratified Okun’s Law analysis; real federal fund rate and yield curve decomposition; M2 velocity and monetary transmission analysis; k-means sub-period temporal clustering; and government debt trajectory with decade-level performance benchmarking. The Keynesian era produced significantly higher mean real GDP growth (3.7% vs. 2.7%), lower structural unemployment (5.0% vs. 6.2%), and a functioning Phillips curve trade-off that collapsed entirely under neoliberalism ($\widehat{\beta }\approx 0$, $p=0.94$). The real federal fund rate averaged negative under the Keynesian regime and turned persistently positive after the Volcker shock. M2 velocity declined sharply in the neoliberal era, refuting the stable quantity theory link central to monetarist policy prescriptions. Temporal clustering recovers five distinct macroeconomic epochs that only partially align with the 1981 historiographical boundary, revealing substantial heterogeneity within each broad regime. The Reagan-era debt expansion, made visible by the corrected debt/GDP trajectory, constitutes a striking empirical contradiction to neoliberal fiscal rhetoric. Together, these results support the interpretation that the two policy regimes represent genuinely distinct macroeconomic equilibria with persistent and measurable consequences for growth, employment, price stability, monetary transmission, and fiscal sustainability. A battery of robustness checks addressing confounding structural shocks, alternative regime boundaries, policy endogeneity, model complexity, and index-construction sensitivity confirms that the core findings are not artifacts of the 1981 periodization, and a cluster-validity analysis formally supports a five-epoch, rather than strictly binary, characterization of the postwar policy landscape. Full article

Wearable inertial sensing and markerless vision are increasingly integrated to enable objective assessment of locomotor and postural function for sports telerehabilitation, intelligent physiotherapy, and athlete performance monitoring. Before such multimodal systems can be translated to clinical practice, their fusion, optimization, and temporal modeling strategies require validation under controlled conditions with reliable ground truth. This study presents a unified multimodal framework that hierarchically integrates inertial measurement unit (IMU) signals and RGB visual information through kernelized representation learning, adaptive multimodal fusion, evolutionary feature optimization, and deep temporal classification. The IMU branch employs Kernelized Extreme Learning Machine (KELM) denoising, Kernelized Canonical Correlation Fusion (KCCF), entropy-guided adaptive windowing, and complementary time-series descriptors (MINIROCKET, TS-CHIEF, and r-STSF). Concurrently, the RGB branch combines anisotropic diffusion filtering, HRNet-based silhouette extraction, DensePose R-CNN, Mesh Graphormer, and Multi-Model Pose-Flow Fusion (MPFF) to learn robust visual representations. Both modalities are integrated through Weighted Canonical Feature Fusion (WCFF) and optimized using a Genetic Algorithm for feature selection and adaptive modality weighting before temporal modeling with cluster-based alignment, Gaussian Process Sequence Modeling, and DeepConvLSTM. As the selected benchmarks do not provide complete inertial recordings, the inertial modality is established according to the adopted experimental protocol to support multimodal fusion analysis. Under 5-fold subject-independent cross-validation, the framework achieves accuracies of 86.56 ± 0.31% on SoccerDiffusion and 88.04 ± 0.25% on HumanoidRobotPose. Although evaluated on humanoid robotic benchmarks, the proposed framework provides a methodological basis for future wearable-enabled clinical movement assessment, remote rehabilitation, and athlete monitoring, while validation on synchronized human inertial-visual datasets remains an important direction for future research. Full article

Temperature fluctuations in cold chain logistics accelerate protein degradation, lipid oxidation, and color deterioration of Litopenaeus vannamei, reducing product value and compromising food safety. To address this issue, this paper proposes a dynamic quality prediction method based on the Tad-Transformer neural network. First, storage experiments were conducted under six cold chain temperature conditions to collect multi-dimensional physicochemical and texture data. Core quality indicators were selected through temperature sensitivity analysis, and a time-series dataset was constructed. Second, an improved K-means++ clustering algorithm incorporating min-max constraints was applied for quality grading. Finally, the Tad-Transformer model was employed to predict quality indicators and temporal grade evolution. Comparative validation with Transformer, Informer and FEDformer on the self-constructed dataset demonstrates that, for the most challenging high-quality samples, the proposed model achieves both precision and recall exceeding 89%, representing improvements of 4.17–10.77% and 2.28–8.37%, respectively, over the comparison models. This method provides technical support for quality grading control and early warning of abnormal risks in cold chain logistics, offering a scientific reference for dynamic quality monitoring and intelligent evaluation of aquatic products. Full article

Ramsey’s RESET test is widely used as an omnibus diagnostic for model misspecification, yet its reliability may be affected by temporal aggregation, autoregressive persistence, and the statistical properties of financial time series. This paper examines the finite-sample performance of the RESET test under temporal aggregation and autoregressive dynamics, with particular reference to exchange-rate modeling. A Monte Carlo simulation framework is developed using linear and benchmark nonlinear autoregressive data-generating processes, alternative persistence parameters, different sample sizes, and aggregation frequencies ranging from daily to annual observations. The simulation evidence is complemented by an empirical application to twelve bilateral exchange-rate return series against the U.S. dollar over the period 1981–2024. Stationarity is assessed using ADF, PP, and KPSS diagnostics, while ARCH–LM diagnostics and White-robust RESET tests are used to evaluate whether RESET rejections may be influenced by conditional heteroskedasticity. The simulations show that the RESET test generally maintains an acceptable size under the null, although it becomes more conservative under higher persistence and lower-frequency aggregation. Under the benchmark polynomial nonlinear specification, the test displays strong power, but the sensitivity analysis shows that power depends on the strength of the nonlinear component and should not be generalized to all nonlinear alternatives. The empirical results indicate frequent RESET rejections at daily and weekly frequencies, while rejection frequencies decline substantially after temporal aggregation. These rejections are interpreted as evidence of general specification error rather than definitive proof of nonlinear conditional-mean dynamics. Robustness tests show that volatility clustering explains part, but not all, of the detected misspecification. Temporal aggregation materially affects RESET test performance and may obscure nonlinear, volatility-related, or other specification features in exchange-rate data. Researchers should therefore interpret RESET results cautiously and consider complementary diagnostics, including dedicated nonlinearity tests, structural-break tests, and models allowing for conditional heteroskedasticity, when modeling persistent or aggregated financial time series. Full article

Under a unified pension system, subnational differences in longevity can translate into implicit pension redistribution through differences in the expected duration of pension receipt. Using Japan as a case study, this paper examines the implicit pension redistribution induced by subnational longevity heterogeneity and analyzes how its measurement varies across spatial scales. Accurate measurement of future redistribution requires forecasts of subnational mortality rates, from which future remaining life expectancy can be derived. Because mortality series at different spatial scales are linked by hierarchical aggregation relationships, independent forecasts may violate aggregation coherence and affect the comparability of redistribution estimates. To address this issue, we introduce and compare several forecast reconciliation methods and select the empirical minimum trace (EMinT) method based on its forecasting performance. Using the reconciled mortality forecasts, we estimate future post-retirement remaining life expectancy across subnational areas and subsequently measure implicit pension redistribution. The results show that subnational longevity heterogeneity will persist throughout the forecast period, generating implicit pension redistribution with clear spatial clustering. Redistribution is substantially greater at the prefectural level than at the regional level, and the difference is projected to widen over time. These findings indicate that analyses conducted at more aggregated spatial scales may underestimate the true extent of pension redistribution across areas. This study provides quantitative evidence for the spatial evaluation of pension systems and public policymaking. Full article

In this work, the structural, optical, vibrational, and stability properties of a series of Al_xAs_x nanostructures are systematically investigated using density functional theory (DFT) and time-dependent density functional theory (TD-DFT). Starting from the fundamental cubic-like Al₄As₄ building block, progressively larger nanostructures were constructed through directional elongation and structural rearrangements, allowing for the exploration of one-dimensional chains, two-dimensional planar structures, and several exotic geometries. The calculated UV–visible absorption spectra reveal that structural dimensionality and topology strongly influence the electronic transitions of the nanostructures, with elongated and distorted configurations exhibiting broader absorption features and richer spectral distribution. Vibrational analysis shows that increasing structural complexity and reducing symmetry lead to a higher density of IR-active modes and more complex infrared spectra. The stability of the nanostructures is evaluated through binding energy calculations, which indicate a clear size-dependent stabilization trend, with the Al₂₄As₂₄-L1 configuration exhibiting the highest stability among the examined systems. In addition, the calculated HOMO-LUMO gaps reveal the semiconducting character of the clusters and demonstrate their sensitivity to geometric topology. The present results establish clear structure–property relationships between dimensional growth and the optical response of AlAs nanoparticles and provide theoretical reference data for future experimental investigations of III-V semiconductor nanostructures. Full article