Search Results (44,251)

The world is undergoing rapid transformations driven by climate change, socio-economic pressures, and geopolitical tensions. Monitoring these dynamics is essential to understand and anticipate territorial change. Although initiatives such as the European Union’s Open Data program promote spatiotemporal datasets (e.g., population, land use), analyzing and interpreting these data over time remains complex and requires technical expertise, limiting their accessibility. This research proposes Semantic Web-based methods to detect and annotate trends in spatiotemporal series, thereby assisting in the systematic analysis of temporal patterns. We introduce the SETT ontology (SEmantic Trajectory of Territory) and its OFF-SETT framework (Ontological Framework For SETT), enabling the formal description of territorial trends and their publication as semantic trajectories in the Linked Open Data cloud. The study delivers (i) a generic methodology for detecting and describing trajectories in spatiotemporal datasets; (ii) a framework for automatically generating knowledge graphs capturing these trajectories; (iii) a knowledge graph describing trajectories of demographic and satellite-derived variables (e.g., temperature, water, vegetation) for study areas in France and Switzerland; and (iv) a web-based geovisualization platform. The approach shows that Semantic Web technologies bridge complex spatiotemporal analysis and public accessibility. By publishing territorial trajectories as knowledge graphs, it fosters transparency, interoperability, and reuse of data, supporting informed decision-making and citizen engagement. Full article

The symmetric Mean Absolute Percentage Error (sMAPE) is a forecast error metric that has been proposed as an alternative to the more common Mean Absolute Percentage Error (MAPE), which is undefined whenever an actual is zero; the sMAPE does not have this problem. Thus, the sMAPE at first glance appears to be more suitable for evaluating forecasts of low volume or intermittent count demand time series. However, the sMAPE suffers from a number of other shortcomings; e.g., it is 2 for a zero actual regardless of the forecast, it always rewards (elicits) integer forecasts 0, 1, 2, ..., if actuals are counts, and it elicits a (typically useless) zero forecast for sufficiently intermittent actuals. This paper collects such properties and discusses their real-world implications so the forecaster can make an informed decision as to whether to use the sMAPE or an alternative. In our opinion, the sMAPE is either unnecessary or dangerous; it should not be used. Full article

Flood risk maps play a critical role in land-use regulation, infrastructure planning, and community preparedness, which are key components of sustainable and climate-resilient urban development. Their production, however, remains costly, labor-intensive, and time-demanding as it relies on simulation-driven workflows that combine hydrodynamic modeling with expert interpretation and extensive validation. To address this issue from a sustainability perspective, we develop a novel, practical, and near-real-time large language model (LLM)-based framework to support property-level flood risk assessment. This framework, which synthesizes geospatial, hydrological, infrastructural, and historical flood information, extends existing research and explores novel risk estimation methods for use in planning practice. Using Brisbane, Australia, as a case study, we develop Flood-LLM, a multi-agent system that transforms multi-source urban datasets into structured textual representations, models diverse neighborhood conditions, and fine-tunes a reasoning model using expert-assessed risk classifications. The results show that Flood-LLM can reproduce official flood risk labels for creek, river, storm tide, and overland-flow hazards with reasonable accuracy, outperforming classical machine learning, deep learning, and untuned LLM baselines. Visual and quantitative analyses indicate that the framework demonstrates a qualitatively nuanced capability to capture salient spatial patterns present in the official maps, while generating a textual chain-of-thought providing a transparent audit trail for its labeling decisions. These findings suggest that such LLM-based approaches can produce potential complementary tools to expert-reviewed planning classifications and support more sustainable, adaptive flood risk management by enabling timely map production and updates that facilitate informed decision-making in rapidly changing environmental conditions. Full article

Accurate hour-ahead forecasting of solar photovoltaic (PV) power is essential for risk-aware decision-making in power systems with increasing renewables. Although recent studies emphasize complex deep learning architectures, it remains unclear whether such complexity provides tangible benefits at very short forecasting horizons, particularly when forecast uncertainty is considered. This study evaluates deterministic and probabilistic hour-ahead PV forecasting using models of varying complexity, including persistence, linear autoregressive models with exogenous inputs, ridge regression, DLinear, and a vanilla long short-term memory (LSTM) network. Probabilistic forecasts were constructed using a unified, model-agnostic, adaptive conformal inference framework incorporating a daily miscoverage reset tailored to the diurnal characteristics of PV generation. Deterministic results indicate that the LSTM achieves the lowest errors, with an RMSE of 0.336 kW (6.55% of rated capacity) and an MAE of 0.164 kW, compared to RMSE values of approximately 0.38–0.45 kW for linear models and persistence. Following conformal calibration, all models attain empirical prediction interval coverage close to the nominal 90% level (PICP ≈ 90.8–91.4%), with performance differences reflected in interval width and sharpness rather than coverage. Notably, linear models combined with adaptive calibration deliver probabilistic performance comparable to the LSTM at substantially lower computational cost. Full article

Background/Objectives: Hemangioblastomas are rare, benign, highly vascular tumors of the central nervous system, frequently associated with von Hippel–Lindau (vHL) disease. Case Presentation: We report a 16-year-old female with vHL presenting with recurrent headaches, abdominal distension, and ocular discomfort. Imaging revealed hemangioblastomas in the fourth ventricle and retrobulbar space, alongside multiple pancreatic cysts. The patient underwent three sessions of Gamma Knife Surgery (GKS) with initial tumor regression and symptom relief. However, long-term follow-up demonstrated progressive disease, with new lesions in the cerebellum, spinal cord, and orbit, including cystic transformation. Histopathology confirmed the reticular variant of hemangioblastoma. Despite further radiosurgical and surgical recommendations, the patient and family opted for conservative management, with lesions remaining radiographically stable over nine years. Conclusions: This case demonstrates that Gamma Knife Surgery may provide temporary local disease control for selected solid hemangioblastomas in von Hippel–Lindau disease but does not alter the underlying disease course. Long-term radiographic stability should be interpreted cautiously, as hemangioblastomas exhibit saltatory growth patterns that make it difficult to distinguish treatment effect from natural tumor quiescence. These findings emphasize that radiosurgery should be regarded as a disease-control strategy rather than curative therapy, underscoring the importance of individualized management, multidisciplinary decision-making, and prolonged surveillance. Full article

The most impactful factors for the cost of fleet management are maintenance expenses and fuel consumption. Traditional ways of monitoring fleet performance fail to connect raw operational data with driving habits. The current study addresses this challenge by developing an architecture of frameworks, consisting of unsupervised and supervised machine learning algorithms, statistical testing, simulation and survival analysis to discover insights that lead to key behavioral predictors. The nucleus of this complex architecture is the decision-making grid (DMG), a two-dimensional matrix that groups vehicles based on their frequency of entering the service and the cost of their repairs. It is the first integration of DMG with ML for prescriptive fleet management. The objective of the study is twofold: firstly, to build a system that classifies vehicles according to their risk profile, and secondly, to offer clear directions for changing driver patterns that most affect vehicle costs or for keeping good practices. The framework proposed by this study not only drives the optimization of operational efficiency but also contributes to a methodology that links driver profiles to costs, offering a scalable methodology for similar business contexts. Full article

This study proposes an explainable machine learning framework for estimating the total project cost (TPC) of AI training-data construction, where cost information is difficult to structure due to heterogeneous workflows and quality requirements. Using 386 public AI training-data projects conducted between 2020 and 2022, we derive 24 numerical predictors from standardized final reports and construct three input tracks: a baseline feature set, a principal component analysis (PCA)-enhanced set, and a factor analysis (FA)–enhanced set capturing latent cost structures. Four regression models (Ridge, Random Forest, XGBoost, and LightGBM) are evaluated using nested cross-validation. XGBoost achieves the best overall performance across all three tracks (Baseline, PCA-enhanced, and FA-enhanced). Among them, PCA-enhanced XGBoost attains the highest predictive accuracy (R² = 0.868; RMSE = 1084.9; MAE = 746.9; MAPE = 0.358; pooled out-of-fold), while Baseline XGBoost yields the lowest MAE (731.4; R² = 0.863). To support transparent decision-making, Shapley Additive exPlanations (SHAP)-based attribution and scenario-based sensitivity analyses are conducted. Results show that project scale and process-level unit costs are dominant cost-drivers, while cloud usage, expert participation, and de-identification requirements exhibit secondary effects. The proposed framework provides an interpretable, data-driven approach to cost information management and decision support for data-intensive AI projects. Full article

The object of this research is resource management and energy consumption processes in optical fiber communication networks with access–metro–core architectures. The study addresses the problem that conventional static and semi-dynamic control methods are unable to simultaneously ensure energy efficiency and QoS stability under conditions of exponentially growing and highly variable traffic. To solve this problem, an AI-based integrated control model was developed that combines traffic prediction, dynamic resource allocation, spectrum management, and power optimization within a unified framework. Traffic prediction is performed using LSTM–BiRNN neural networks (1.2–1.8 million parameters, 300–500 thousand records), while control decisions are generated by an Actor–Critic reinforcement learning algorithm. Simulation results obtained in the Python 3.12 and OptiSystem 17.0 environments demonstrate that, in the Access segment (1–10 Gb/s), latency is stabilized within 1–10 ms; in the Metro segment (40–120 Gb/s), energy consumption is reduced by 18–27%; and in the Core segment (400–1000 Gb/s), the efficiency of RSA algorithms increases by 22–35%. When the EDFA output power is maintained within +17 to +23 dBm, amplifier power consumption decreases by 10–15%, resulting in overall network energy savings of 20–40%. The obtained results are explained by the synergy of accurate traffic prediction provided by the LSTM–BiRNN model and proactive real-time decision-making enabled by the Actor–Critic algorithm. The distinctive feature of the proposed approach is the simultaneous optimization of energy efficiency and QoS across all access, metro, and core segments within a single integrated architecture. The results can be practically applied in the design and modernization of optical fiber communication networks, as well as in the deployment of energy-efficient intelligent network management systems. Full article

This study investigates the relationship between digital financial literacy (DFL) and hyperbolic discounting among 104,993 active securities account holders in Japan. As digital financial services expand rapidly, individuals increasingly require not only traditional financial knowledge but also the capacity to understand digital platforms, evaluate online financial information, and manage emerging technological risks. Using data from the 2025 wave of the Survey on Life and Money, hyperbolic discounting is measured through intertemporal monetary choice scenarios, while DFL is constructed as a multidimensional index encompassing digital knowledge, financial knowledge, service awareness, attitudes, behaviors, practical capability, and protection against digital fraud. Probit regression results reveal a statistically significant negative association between DFL and hyperbolic discounting, indicating that individuals with stronger digital financial competencies are less likely to exhibit hyperbolic discounting. Attitudinal components of DFL exhibit the strongest effects, suggesting that internalized financial beliefs may play a more decisive role than technical knowledge in promoting time-consistent decision-making. Subsample analyses further highlight gender-differentiated patterns in demographic and economic influences on present bias. These findings contribute to behavioral finance by integrating digital capability into intertemporal choice research and provide policy-relevant implications for designing comprehensive financial education and digital literacy initiatives in increasingly digitalized financial environments. Full article

Legal decision-making is expected to meet high standards of consistency and rationality, yet human judgments in this domain are known to be influenced by procedural factors such as evidence order and intermediate evaluations. Recent work has shown that even legal professionals, including judges, are susceptible to such biases when assessing criminal cases. This raises a critical question: do large language models, which are increasingly proposed as decision-support tools in legal contexts, exhibit similar procedural biases—and if so, can these biases be mitigated? To address this question, we tested GPT-4o and GPT-5.2 using a controlled legal judgment task adapted from prior human research. The task involved simplified criminal cases in which we systematically manipulated (i) the order of incriminating and exonerating evidence and (ii) whether an intermediate guilt judgment was required before a final decision. Model responses were directly compared to human judgments from the original study. We additionally examined whether prompt engineering strategies, based on current best-practice recommendations, could reduce observed biases. GPT-4o exhibited robust order effects and a form of evaluation bias, although the latter differed in structure from the human pattern. GPT-5.2 showed similar but attenuated effects. Across both models, prompt engineering had limited and inconsistent impact, failing to reliably eliminate procedural sensitivity. These findings suggest that even advanced large language models remain vulnerable to normatively irrelevant procedural influences. More broadly, they advise caution in treating large language models as inherently rational or bias-resistant decision-support systems in high-stakes professional domains such as law. Full article

Genetically modified foods represent an important application of modern biotechnology and remain a subject of public debate. Attitudes toward consumption are more likely to be influenced by varying levels of scientific knowledge. University students from the College of Science and the College of Agricultural and Food Sciences at King Faisal University, Saudi Arabia, are expected to possess the basic knowledge that may affect their attitudes toward consumption of genetically modified foods. This study aimed to assess undergraduate students’ knowledge as a predictor of future attitudes toward consumption of genetically modified foods. Using a descriptive method, an electronic questionnaire was administered to a random sample of 300 participants during the first semester of the academic year 2025/2026. Data were analyzed using confirmatory factor analysis and t-tests. The results indicate that students possess a moderate level of scientific knowledge. Their future attitudes toward consuming genetically modified foods were also moderate. Prior studying of genetics and biotechnology courses significantly affects students’ scientific knowledge and future attitudes toward consumption of genetically modified food. The students strongly supported strict regulations, but they expressed hesitation regarding consumption regardless of scientific assurances of safety. No statistically significant differences in knowledge or attitudes based on specialization or gender were found. The authors recommend integrating ethical and social considerations of this kind of food into educational curricula to support informed decision-making among future professionals. Full article

Introduction: In patients with conductive and mixed hearing loss, implantable hearing devices such as active middle ear implants (AMEIs) and bone conduction implants (BCIs) are established alternatives when conventional hearing aids fail. Although bone conduction (BC) thresholds are routinely used as eligibility criteria, their role as frequency-specific predictors of postoperative functional outcomes remains poorly defined. This study aimed to evaluate the influence of preoperative BC thresholds across the audiometric spectrum on postoperative speech recognition outcomes after implantation with AMEIs and BCIs. Methods: A retrospective observational study was conducted at a tertiary referral center including patients implanted with BCIs or AMEIs. Pre- and postoperative audiological data were analyzed, including air and bone conduction thresholds, frequency-segmented BC measures (low, mid, and high frequencies), cochlear frequency gradient (ΔBC Slope), and speech recognition scores (SRSs) at 65 dB HL one year after implantation. Results: 102 patients were included (50 BCI, 52 AMEI). Both implant types achieved significant postoperative improvements in tonal thresholds and SRS compared with pre-implantation values (all p < 0.001). High-frequency BC thresholds (BC-High, 4–6 kHz) showed a significant inverse correlation with postoperative SRS in both BCI (r = −0.382, p = 0.001) and AMEI users (r = −0.398, p < 0.001), and emerged as the only independent predictor in multivariable models (BCI: β = −0.533, p = 0.022; AMEI: β = −0.491, p = 0.020). Low- and mid-frequency BC measures were not associated with postoperative speech outcomes (all p > 0.05). ROC analyses demonstrated excellent discriminative performance of BC-High for identifying suboptimal outcomes, with area under the curve values of 0.92 for BCI (p = 0.001) and 0.94 for AMEI (p = 0.002), and implant-specific cutoff values of >47 dB HL and >61 dB HL, respectively. Conclusions: High-frequency BC thresholds showed the strongest association with postoperative speech recognition after implantable hearing rehabilitation. BC-High could function as a prognostic marker of functional outcome rather than an eligibility criterion, providing clinically meaningful information to refine preoperative counseling and individualized decision-making within current indication frameworks. Full article

The combination of Artificial Intelligence (AI) and Formative Assessment (FA) in Teacher Education explores how emerging technologies can enhance teaching practices and professional development. AI tools can provide personalized feedback, identify learning needs, and support reflective practice among educators. Integrating AI-driven formative assessment methods allows for continuous evaluation of teaching competencies, promoting adaptive learning, data-informed decision-making, and improved instructional quality in teacher education programs. The purpose of this study was to conduct a systematic review of the use of Formative Assessment (FA) and Artificial Intelligence (AI) in Teacher Education (TE) during the period 2020–2025 (inclusive). The review was conducted following the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) methodology, which ensures a rigorous, transparent, and reproducible process in the selection and analysis of studies. To this end, scientific articles published in the Scopus, Web of Science and Dialnet databases were reviewed, considering publications in English and Spanish. The objective was to identify trends, methodological approaches, results, and research gaps that show how AI is being integrated, or not, into FA processes in TE. The review also sought to analyze the impact of AI on student participation in assessment, feedback, decision-making, and the learning and assessment process itself, synthesizing the current evidence on the relationship between AI and FA in TE. Full article

The smart city represents a new stage in urban evolution, driven by technological progress, social transformations, and the increasing emphasis placed on sustainability. This metamorphosis generates hub-type architectural models, used not only for data collection and interconnection but also for the management and monitoring of people, resources, and urban services. This discussion addresses how digital urbanism has followed different paths globally by synthesising technological, economic, social, and governance perspectives. Compared with traditional models of urbanisation, new smart cities are built not only for digital interconnection but also to be citizen-centred, environmentally friendly, and resilient to global crises. This article analyses recent scientific literature on the theoretical and practical foundations of technologies that support data-driven decision-making, infrastructure efficiency, and the delivery of inclusive public services. At the same time, major challenges are highlighted, such as the lack of system interoperability, information fragmentation, and the risks associated with excessive surveillance, which can generate social exclusion, as well as financial and political constraints. International examples from Helsinki, Barcelona, Dubai, and Singapore offer both models that have achieved success and critical lessons about the limits of these approaches. This paper is not limited only to the problems faced by smart cities. It also highlights the opportunities they can bring. Finally, based on the conclusions of the analysis carried out and the identified trends, a strategic framework is proposed, oriented towards responsible innovation, collaboration, and sustainability. This approach contributes to informing researchers, decision-makers, urban planners, and the public interested in the transformation of the urban environment. Full article

The issue of natural protected areas in the urban environment is not a new topic at the European level, but its approach differs from one state to another, depending on phrasing the issues in a particular context. Romania was selected as study area because, despite its exceptionally rich natural heritage, no urban-planning instruments dedicated to the areas adjacent to natural protected sites have been adopted so far. The purpose of this article is to identify what kind of tools can be adopted for a two-way support—both with respect to planning and the natural heritage. The key roles of areas adjacent to urban natural protected sites are identified in order to establish appropriate descriptive terms. The principles and objectives required for planning these zones are critically examined, enabling an assessment of their applicability and quantifying their potential through urban indicators, indices, and specific planning measures. Ultimately, following the formulation supporting instruments, the study highlights the need for an adapted urban-planning documentation structure tailored to such sensitive territories and the need to provide public access to information through a dedicated platform supporting informed decision-making. Full article