Search Results (34,208)

Crude oil price forecasting contributes to global economic development. This study proposes a hybrid deep learning model for crude oil price forecasting. First, empirical wavelet transform decomposes raw data into multiple. Then, three neural networks generate preliminary forecasts, which are subsequently refined by a reinforcement learning-based ensemble method. Finally, an error correction module handles residuals, further enhancing the forecasting outcomes. Three West Texas Intermediate datasets and additional emergency scenarios were used to validate the hybrid model. The findings indicate that the proposed model achieves superior predictive performance compared with sixteen benchmark methods and three advanced models. Full article

The phenomenon of patient mobility is becoming increasingly frequent, altering the actual service ranges of hospitals across various cities. However, its impact on the spatial equity of healthcare services at the national scale has yet to be fully explored. This paper aims to reveal the impact of intercity patient mobility on healthcare equity in China. Using one million patient mobility records from online healthcare platforms, we construct the 2023 Cross-City Patient Mobility Network in China and identify the patterns of cross-city patient mobility. Furthermore, we employ the Dagum Gini coefficient to measure the spatial disparities in per capita healthcare services before and after patient mobility. The results show that: (1) cross-city patient mobility exhibits administrative boundary effects and reflects the administrative hierarchy system, yet megacities extend their healthcare service ranges beyond provincial and urban agglomeration boundaries; (2) patient mobility enhances the equity of per capita healthcare services at both intra-provincial and inter-provincial levels, with inter-provincial disparities contributing significantly more than intra-provincial disparities—a trend further reinforced by patient mobility. This study not only provides a methodological framework for understanding the impact of patient mobility on the healthcare system but also offers empirical support for public health policymaking. Full article

Cross-Domain Few-Shot Learning (CD-FSL) aims to empower machines with the capability to rapidly acquire new concepts across domains using an extremely limited number of training samples from the target domain. This ability hinges on the model’s capacity to extract and transfer generalizable knowledge from a source training set. Studies have indicated that the similarity between source and target-data distributions, as well as the difficulty of target tasks, determine the classification performance of the model. However, the current lack of quantitative metrics hampers researchers’ ability to devise appropriate learning strategies, leading to a fragmented understanding of the field. To address this issue, we propose quantitative metrics of domain distance and target difficulty, which allow us to categorize target tasks into three regions on a two-dimensional plane: near-domain tasks, far-domain low-difficulty tasks, and far-domain high-difficulty tasks. For datasets in different regions, we propose a Divide-and-Conquer Strategy (DCS) to tackle few-shot classification across various target datasets. Empirical results across 15 target datasets demonstrate the compatibility and effectiveness of our approach, improving the model performance. We conclude that the proposed metrics are reliable and the Divide-and-Conquer Strategy is effective, offering valuable insights and serving as a reference for future research on CD-FSL. Full article

This study examines users’ perceptions of Chatbots in China, with a particular focus on the factors influencing their acceptance and usage. Grounded in the Technology Acceptance Model (TAM), we analyze data from 542 online responses to explore the roles of Perceived Authenticity, usefulness, and risk in shaping user behavior toward AI-driven Chatbots. Using linear regression and mediation analyses, our findings indicate that both Perceived Authenticity and Perceived Usefulness positively impact users’ behavioral intentions, while Perceived Risk has a negative influence. Notably, Perceived Usefulness serves as a mediator between behavioral intentions and both Perceived Authenticity and Perceived Risk. These results contribute to the growing body of research on AI and e-commerce by providing empirical evidence of the key factors affecting Chatbot adoption. The study offers valuable implications for developers and marketers, suggesting that enhancing Perceived Authenticity and usefulness while addressing Perceived Risks can improve user acceptance. These insights are particularly pertinent for AI practitioners aiming to refine Chatbot technology and expand its application across various sectors. Full article

In this study, we introduce a detailed framework for predicting market conditions and selecting stocks by integrating machine learning techniques with ontological financial analysis. The process starts with ontology-based stock selection, categorizing companies using fundamental financial indicators such as liquidity, profitability, debt ratios, and growth metrics. For instance, firms showcasing favorable debt-to-equity ratios along with robust revenue growth are identified as high-performing entities. This classification facilitates targeted analyses of market dynamics. To predict market states—categorizing them into bull, bear, or neutral phases—the framework utilizes a Non-Stationary Markov Chain (NMC), BERT, to assess sentiment in financial news articles and Long Short-Term Memory (LSTM) networks to identify temporal patterns. Key inputs like the Sentiment Index (SI) and Illiquidity Index (ILLIQ) play essential roles in dynamically influencing regime predictions within the NMC model; these inputs are supplemented by variables including GARCH volatility and VIX to enhance predictive precision further still. Empirical findings demonstrate that our approach achieves an impressive 97.20% accuracy rate for classifying market states, significantly surpassing traditional methods like Naive Bayes, Logistic Regression, KNN, Decision Tree, ANN, Random Forest, and XGBoost. The state-predicted strategy leverages this framework to dynamically adjust portfolio positions based on projected market conditions. It prioritizes growth-oriented assets during bull markets, defensive assets in bear markets, and maintains balanced portfolios in neutral states. Comparative testing showed that this approach achieved an average cumulative return of 13.67%, outperforming the Buy and Hold method’s return of 8.62%. Specifically, for the S&P 500 index, returns were recorded at 6.36% compared with just a 1.08% gain from Buy and Hold strategies alone. These results underscore the robustness of our framework and its potential advantages for improving decision-making within quantitative trading environments as well as asset selection processes. Full article

Urinary tract infections caused by Gram-negative bacteria (GNB) are among the most common infections and a significant cause of sepsis. The increasing prevalence of multidrug-resistant (MDR) bacteria poses challenges to empirical treatment. Colistin may be used a last-resort antibiotic for treating MDR infections, but this requires the rapid determination of susceptibility to colistin. Traditional susceptibility testing methods can take up to 48 h, and there are specific challenges in determining colistin susceptibility. This study evaluates a novel, rapid method for determining colistin susceptibility directly from positive urine samples using the FASTcolistin MIC kit from FASTinov^®. A total of 100 urine samples positive for Gram-negative bacilli when screened by the UF-1000i system were included in this study. After a simple sample prep, the same bacterial suspension was used for identification on MALDI-TOF and inoculated in the FASTcolistin MIC panel for our AST; after incubation at 37 °C for 1 h, it was analyzed via flow cytometry using a CytoFLEX cytometer (Beckman Coulter, Brea, CA, USA). The categorical susceptibility to colistin according to EUCAST or CLSI standards as well as the MIC values were given by bioFAST software (bioFAST 2.0). The essential agreement (EA) and bias were calculated. Different species of Enterobacterales, Pseudomonas aeruginosa, and Acinetobacter spp. were correctly identified by MALDI-TOF directly from the FASTcolistin MIC sample prep. The essential agreement between the two methods was 99%, with a bias of −17%. Both identification and susceptibility were obtained in less than 2 h. This study presents a rapid and accurate method for determining colistin MIC directly from urine samples. The shortness of time required to produce a result, 2 h versus 48 h with the conventional methods, will significantly impact treatment decisions, especially in urinary tract infections difficult to treat. Full article

This study presents a novel approach to analyzing forest regeneration dynamics by integrating a Markov chain model with Multivariate Time Series (MTY) decomposition. The probabilistic tracking of age-class transitions was combined with the decomposition of regeneration rates into trend, seasonal, and irregular components, unlike traditional deterministic models, capturing the variability and uncertainties inherent in forest ecosystems, offering a more nuanced understanding of how Scots pine (Pinus sylvestris L.) and other tree species evolve under different management and climate scenarios. Using 20 years of empirical data from the Lithuanian National Forest Inventory, the study evaluates key growth and mortality parameters for Scots pine, Spruce (Picea abies), Birch (Betula pendula), and Aspen (Populus tremula). The model for Scots pine showed a 79.6% probability of advancing from the 1–10 age class to the 11–20 age class, with subsequent transitions of 82.9% and 84.1% for older age classes. The model for Birch shown a strong early growth rate, with an 84% chance of transitioning to the next age class, while the model for Aspen indicated strong slowdown after 31 years. The model indicated moderate early growth for Spruce with a high transition in later stages, highlighting its resilience in mature forest ecosystems. Sensitivity analysis revealed that while higher growth rates can prolong forest stand longevity, mortality rates above 0.33 severely compromise stand viability. The Hotelling $T^2$ control chart identified critical deviations in forest dynamics, particularly in years 13 and 19, suggesting periods of environmental stress. The model offers actionable insights for sustainable forest management, emphasizing the importance of species-specific strategies, adaptive interventions, and the integration of climate change resilience into long-term forest planning. Full article

The problem of maximizing energy efficiency in Algerian residential structures in Mediterranean climates is discussed in this article. The primary issue with North Africa’s residential building stock is the dearth of benchmark models that describe thermal comfort and energy use, which is made worse by high cooling needs and energy poverty. The principal aim of this study is to create a benchmark model that will aid in evaluating the energy performance of the existing system and to suggest a series of actions to improve efficiency and thermal comfort in the future. The technique builds a calibrated model based on a database of 284 Algiers apartments by combining modeling and empirical observations. Based on the observed U-Value wall of 0.43 W/(m²K), the average annual energy use for Archetype A is 3.70 kWh/m², and the average annual heating energy use is 13.20 kWh/m². The significance of this model in advancing energy efficiency and sustainability in Mediterranean climates is emphasized in the Conclusion Section. These results contribute to our understanding of the dynamics of building energy in similar global environments and evaluate the thermal comfort and the measurement of CO₂ emissions in this type of building. Full article

Historically, urban development has always been centered on coastal areas, with access to waterbodies—seas, rivers and canals—being a significant advantage for movement and trade. With most of the world’s megapolises located on coasts, land reclamation offers a solution for the expansion of city centers which are otherwise restricted by the coastline. This study aims to define the current understanding of urban regeneration and development on reclaimed lands, addressing the basic questions of what, why and how. This study aims to assess urban regeneration on reclaimed coastal land based on the principles of sustainable development defined by existing studies. The literature review establishes a theoretical framework and defines performance-based benchmarks for identifying spatial indicators of urban development. Composite indicators, namely open space coverage, land use mix, the percentage of coast for people, accessibility to public transportation and amenities, the availability of pedestrian paths and cycling tracks and adequate road networks, are considered for this framework. The conclusions are drawn based on the results of an analysis of spatial layout using a GIS as a tool to map and empirically measure each indicator. The framework is validated using a major land reclamation project, West Bay, in the coastal urban area of Doha in Qatar. The results determine that West Bay has achieved a good level of sustainability, although there are areas that could be enhanced to improve the overall sustainability of urban development further. These findings can serve as a guide for policymakers and various stakeholders for sustainable urban planning on reclaimed coastal lands. Full article

With the rapid expansion of urban metro networks, metro stations, as critical public spaces, not only influence transit efficiency and spatial layout optimization but also play a vital role in shaping users’ emotional experiences. This study examines metro station spaces as its primary research subject, exploring the correlation between physical environmental features and emotional perception within the framework of environmental psychology theory. This study adopts an innovative approach by integrating deep learning-based affective computing methods with semantic segmentation techniques in computer vision to systematically evaluate the impact of various physical environmental features and functional spaces on users’ emotional perceptions across multiple dimensions. The study provides empirical evidence for assessing and interpreting the relationship between environmental features and emotional perception, thereby enhancing the reliability of the research. The findings, quantified through deep learning methods, identify key factors influencing various emotional perception scores in metro stations. These insights will assist practitioners in gaining a deeper understanding of how metro station spaces impact users’ emotional experiences and can be applied to early-stage design and later-stage optimization of metro station spaces. Full article

Accurate estimation of concrete compressive strength is very important for the improvement of mix design, quality assurance, and compliance with engineering specifications. Most empirical traditional models have failed to capture the complex relationships inherent within varied constituents of concrete mixes. This paper develops a machine learning model for compressive strength prediction using mix design variables and curing age from a “Concrete Compressive Strength Dataset” obtained from the UCI Machine Learning Repository. After comprehensive data preprocessing and feature engineering, various regression and classification models were trained and evaluated, including gradient boosting, random forest, AdaBoost, k-nearest neighbors, linear regression, and neural networks. The gradient boosting regressor (GBR) achieved the highest predictive accuracy with an R² value of 0.94. Feature importance analysis showed that the water–cement ratio and age are the most crucial factors affecting compressive strength. Advanced methods such as SHapley Additive exPlanations (SHAP) values and partial dependence plots were used to attain deep insights about feature interaction with a view to enhancing interpretability and fostering trust in models. Results highlight the potential of machine learning models to improve concrete mix design with the aim of sustainable construction through the optimization of material usage and waste reduction. It is recommended that future research be undertaken with expanding datasets, more features, and richer feature engineering to enhance predictive power. Full article

In human dynamics, functioning relies on intricate coordination patterns. Networks of synchronized oscillators in various biological and semiotic fields shape these dynamics. We have observed stability, instability, and transitions at multiple levels, indicating that coordination happens on all scales. We have examined coordination models for simplified and complex dynamics. In empirical research, we can frequently observe chimera states as the coexistence of coherence and incoherence, even in homogeneous networks. They are more evident in the heterogenous networks’ standard in human dynamics, where oscillators and nodes are mixed as different types. This paper proposes a simplified and overarching model for mixed chimeras. We discuss the information dynamics in these types of networks and their pattern transitions. Full article

Text classifiers are Artificial Intelligence (AI) models used to classify new documents or text vectors into predefined classes. They are typically built using supervised learning algorithms and labelled datasets. Text classifiers produce a predefined class as an output, which also makes them susceptible to adversarial attacks. Text classifiers with high accuracy that are trained using complex deep learning algorithms are equally susceptible to adversarial examples, due to subtle differences that are indiscernible to human experts. Recent work in this space is mostly focused on improving adversarial robustness and adversarial example detection, instead of detecting adversarial attacks. In this paper, we propose a novel approach, explainable AI with integrated gradients (IGs) for the detection of adversarial attacks on text classifiers. This approach uses IGs to unpack model behavior and identify terms that positively and negatively influence the target prediction. Instead of random substitution of words in the input, we select the top p% words with the greatest positive and negative influence as substitute candidates using attribution scores obtained from IGs to generate k samples of transformed inputs by replacing them with synonyms. This approach does not require changes to the model architecture or the training algorithm. The approach was empirically evaluated on three benchmark datasets, IMDB, SST-2, and AG News. Our approach outperforms baseline models on word substitution rate, detection accuracy, and F1 scores while maintaining equivalent detection performance against adversarial attacks. Full article

This paper examines the intertwined dynamics among digital transformation, IT innovation, and sustainability and their collective influence on firm performance in response to the evolving business landscape characterized by digitalization, IT innovation, and sustainability concerns. The study investigates how these factors collectively impact firm performance by analyzing a panel dataset of 1510 global companies from 2013–2023. The model utilizes a multiple linear regression analysis to incorporate firm performance scores as the dependent variable. At the same time, digital transformation, IT innovation, and sustainability factors are the independent variables, alongside firm-level control variables. The results reveal that digital transformation positively influences IT innovation and strategic business model (BM) development, confirming its direct impact on firm performance. Additionally, firms with simpler and younger structures achieve better outcomes than larger and more established ones. However, the study has limitations, as it is based on a panel dataset spanning 11 years; extending the analysis to a different and longer period could provide insights into the evolving nature of digital transformation, which is inherently dynamic. This study is groundbreaking in exploring these factors, offering a unique perspective through its analysis of an 11-year panel and its focus on assessing dynamic business models. Full article

The main objective of this paper is to examine the impact of ESG management on corporate performance by focusing on board characteristics. To this end, this study uses financial data and empirical panel data of Fortune 300 firms from 2008 to 2021 and firm-specific ESG scores derived from the European Sustainability Reporting Standard (ESRS) to conduct an empirical analysis. Specifically, a panel model analysis was conducted to examine the relationship between ESG management and firm performance using alternative variables on board characteristics. In the basic model analysis, we adopted alternative variables for ESG management and board characteristics and conducted a panel model analysis to examine the relationship between these factors and corporate performance. In the basic model analysis that included board characteristics, only board size (+) and nationality diversity (−) had a statistically significant effect on corporate performance, while gender diversity had no statistically significant effect on corporate performance. However, in the full model analysis, where board characteristics and ESG management were combined, factors E (−) and S (+) had statistically significant effects on firm performance, confirming that the presence of a board of directors leads to better performance. We found that the effects of E and S on firm performance were reversed, indicating that there is a difference in the cost of ESG management by factor. Finally, G did not have a statistically significant relationship with firm performance, which was likely due to the fact that the characteristics of the board were already reflected in ESG, confirming the role of the board. As a result, the board of directors seems to help with the smooth implementation of ESG management by focusing on internal stabilization and communication, suggesting that future research should consider the impact of the board of directors rather than analyzing ESG management in isolation. The results also show that the board of directors in the G sector has a significant impact on ESG management, but it is not treated as an important factor in ESG evaluation criteria, suggesting that it is necessary to reflect factors on stakeholder communication. Finally, the practical implication is that a united board is necessary to implement ESG management in corporate operations. Full article