This study investigates factors influencing employees' perceptions of the usefulness of Business Process Management Systems (BPMS) in commercial settings. It explores the roles of system dependency... Show More

This study investigates factors influencing employees' perceptions of the usefulness of Business Process Management Systems (BPMS) in commercial settings. It explores the roles of system dependency, system quality, and the quality of information and knowledge in the adoption and use of BPMS. Data were collected using a structured questionnaire from end-users in various firms and analyzed with Partial Least Squares (PLS). The survey evaluated perceptions of service quality, input quality, system attributes, and overall system quality. The findings indicate that service quality, input quality, and specific system attributes significantly influence perceived system quality, while system dependency and information quality are predictors of perceived usefulness. The results highlight the importance of user training, support, and high-quality information in enhancing satisfaction and BPMS. This research offers empirical evidence on the factors impacting user perceptions and acceptance, emphasizing the need for user-centric approaches in BPMS.

Risk prediction has become increasingly crucial in today's complex and dynamic environments. However, existing forecasting methods still face challenges in terms of accuracy and reliability.... Show More

Risk prediction has become increasingly crucial in today's complex and dynamic environments. However, existing forecasting methods still face challenges in terms of accuracy and reliability. Therefore, it is imperative to explore new approaches to better address risks. In response to this need, our study introduces an innovative risk prediction model known as WOA-FPALSTM. What sets this model apart is its seamless integration of deep learning and heuristic algorithms, designed to overcome the limitations of existing approaches. The core component of deep learning, LSTM, excels in sequence modeling by capturing long-term and short-term dependencies in time series data, thereby enhancing the model's ability to model temporal data. Meanwhile, the heuristic algorithm, WOA (Whale Optimization Algorithm), equips our model with global search capabilities, facilitating the discovery of optimal model configurations and significantly improving predictive performance.

This study builds a research model based on sense-making theory and Dervin's sense-making model of 'gap-bridge-uses' to explore the relationship between users' fear of missing out (FoMO) in relation... Show More

This study builds a research model based on sense-making theory and Dervin's sense-making model of 'gap-bridge-uses' to explore the relationship between users' fear of missing out (FoMO) in relation to charity crowdfunding performance, and collected a set of 230 data of users from the famous charitable crowdfunding campaign on TikTok. The research revealed that FoMO could predict approach behaviour towards charitable crowdfunding. And collective psychological ownership (CPO) mediated the effect of FoMO on approach behaviour towards charitable crowdfunding, and psychological resilience (PR) positively moderated the mediating effect of CPO on approach behaviour. This study not only enriched the research about information interaction and sense-making process between users and charitable crowdfunding, but also extended the research boundary of the relationship between FoMO and approach behaviour towards charitable crowdfunding.

This paper proposes a smart learning system built on deep learning and augmented reality (AR) to support employees with practical IoT experimentation, from components and circuit board pin... Show More

This paper proposes a smart learning system built on deep learning and augmented reality (AR) to support employees with practical IoT experimentation, from components and circuit board pin connections to programming and IoT control. For instance, employees can use their mobile phones to capture images of electronic components and access AR-enhanced instructional materials for learning component properties. This AR-assisted learning system offers guidance at each experimental stage to support hands-on practice and troubleshooting. The learning system also incorporates the pair programming teaching method to enhance experimental quality and confidence, enabling employees to collaborate with teammates throughout the process. The system is further equipped with an online whiteboard for Q&A and in-depth theoretical exploration at each stage of the experiment. Additionally, a blockchain platform records and analyzes each employee's learning progress and status, providing a comprehensive view of their development.

This paper proposes an emotion-aware personalized recommendation system (EPR-IoT) based on IoT data and multimodal emotion fusion, aiming to address the limitations of traditional recommendation... Show More

This paper proposes an emotion-aware personalized recommendation system (EPR-IoT) based on IoT data and multimodal emotion fusion, aiming to address the limitations of traditional recommendation systems in capturing users' emotional states of artistic product consumption in real time. With the proliferation of smart devices, physiological signals such as heart rate and skin conductance—which are strongly correlated with emotional states—provide new opportunities for emotion recognition. For example, an increase in heart rate is typically associated with emotions like anxiety, anger, or fear, while a decrease is linked to emotional states like relaxation or joy. Similarly, skin conductance rises with emotional arousal, particularly during stress or fear. These physiological signals, combined with text, speech, and video data of art products, are fused to construct an art emotion-driven recommendation model capable of dynamically adjusting the recommended content.

In domains such as e-commerce and media recommendations, personalized recommendation systems effectively alleviate the issue of information overload. However, existing systems still face challenges... Show More

In domains such as e-commerce and media recommendations, personalized recommendation systems effectively alleviate the issue of information overload. However, existing systems still face challenges in multimodal data processing, data sparsity, and dynamic changes in user preferences. This paper proposes a Hierarchical Generative Reinforcement Learning Recommendation Optimization framework (HG-RLRO) that addresses these issues by integrating multimodal data, Generative Adversarial Networks (GAN), Inverse Reinforcement Learning (IRL), and Hierarchical Temporal Difference Learning (HTD). HG-RLRO employs a multi-agent architecture to handle textual and image data and utilizes GAN to generate simulated user behavior data to mitigate data sparsity. IRL dynamically infers user preferences across multiple time scales.

In response to the challenges posed by globalization and rapid technological advancements, traditional static pricing models are no longer sufficient to capture the dynamic nature of consumer... Show More

In response to the challenges posed by globalization and rapid technological advancements, traditional static pricing models are no longer sufficient to capture the dynamic nature of consumer behavior and market fluctuations. This study proposes a “Multi-dimensional Dynamic Pricing Optimization and Consumer Behavior Prediction Model Driven by Big Data,” which integrates multi-source data and reinforcement learning to improve dynamic pricing strategies. Through a hybrid model architecture using Random Forest and LSTM, it captures both static and time-series features. Experimental results show that the proposed model significantly outperforms baseline models, achieving a 43% reduction in Mean Squared Error (MSE), a 28% decrease in Mean Absolute Percentage Error (MAPE), a 6.5% increase in Accuracy, and a 14.7% increase in Cumulative Revenue. These findings confirm the model's ability to enhance prediction accuracy, optimize pricing strategies, and maximize revenue, demonstrating its potential for real-world applications in industries like e-commerce, finance, and advertising.

In the era of the digital economy in which disruptive information technologies such as artificial intelligence, blockchain, big data, and cloud computing prevail, digital marketing is increasingly... Show More

In the era of the digital economy in which disruptive information technologies such as artificial intelligence, blockchain, big data, and cloud computing prevail, digital marketing is increasingly becoming the focus of common attention in practice and academia. In this paper, companies listed on the GEM from 2019 to 2023 are selected as research samples to empirically test the impact and intrinsic role of digital marketing on enterprise performance. The results show that digital marketing significantly stimulates enterprise performance, and its incentive effect is mainly reflected in strategic innovation rather than substantive innovation. It is still valid after a series of endogeneity and robustness tests such as instrumental variable method and propensity score matching method. Based on the new perspective of information dynamic capability, this study opens the “black box” of digital marketing enabling enterprise performance and providing certain enlightenment for enterprises to effectively realizing the “win-win” of digital transformation and innovation ability improvement.

Predictive maintenance is gaining increasing attention in the field of industrial equipment management as an effective strategy to enhance equipment reliability and reduce maintenance costs. Deep... Show More

Predictive maintenance is gaining increasing attention in the field of industrial equipment management as an effective strategy to enhance equipment reliability and reduce maintenance costs. Deep learning has become a focal point due to its exceptional ability to process time series data and recognize complex patterns. To address challenges related to accuracy and robustness in predicting equipment failures, this study proposes a novel model that combines deep reinforcement learning (DDPG) with gated recurrent units (GRU), alongside Bayesian Optimization for hyperparameter tuning. The DDPG component learns the dynamic interactions between actions and states, adapting to the specific characteristics of different devices. The GRU module is designed to capture temporal dependencies in sensor data.

As concerns over environmental pollution and the reduction of greenhouse gas emissions intensify, sustainable strategies in supply chain transportation are critical. This paper proposes a novel... Show More

As concerns over environmental pollution and the reduction of greenhouse gas emissions intensify, sustainable strategies in supply chain transportation are critical. This paper proposes a novel approach to optimizing transportation routes and reducing carbon emissions in a green supply chain using deep reinforcement learning. The research targets a three-tier green supply chain consisting of manufacturers, third-party logistics providers (3PL), and retailers. First, a carbon reduction model for transportation is established, accounting for both product greenness and carbon emissions that influence demand. The study then introduces a Proximal Policy Optimization (PPO)-based contract model, combining cost-sharing and profit-sharing mechanisms between retailers and logistics providers to incentivize eco-friendly practices.

Against the backdrop of increasingly severe global environmental changes, accurately predicting and meeting renewable energy demands has become a key challenge for sustainable business development.... Show More

Against the backdrop of increasingly severe global environmental changes, accurately predicting and meeting renewable energy demands has become a key challenge for sustainable business development. Traditional energy demand forecasting methods often struggle with complex data processing and low prediction accuracy. To address these issues, this paper introduces a novel approach that combines deep learning techniques with environmental decision support systems. The model integrates advanced deep learning techniques, including LSTM and Transformer, and PSO algorithm for parameter optimization, significantly enhancing predictive performance and practical applicability. Results show that our model achieves substantial improvements across various metrics, including a 30% reduction in MAE, a 20% decrease in MAPE, a 25% drop in RMSE, and a 35% decline in MSE. These results validate the model's effectiveness and reliability in renewable energy demand forecasting. This research provides valuable insights for applying deep learning in environmental decision support systems.

With the growing severity of global climate change, achieving carbon neutrality has become a central focus worldwide. The intersection of population studies and carbon neutrality introduces... Show More

With the growing severity of global climate change, achieving carbon neutrality has become a central focus worldwide. The intersection of population studies and carbon neutrality introduces significant challenges in predicting and optimizing energy consumption, as demographic factors play a crucial role in shaping carbon emissions. This paper proposes a model based on a Region-based Convolutional Neural Network (RCNN) and Generative Adversarial Network (GAN), enhanced with a dual-stage attention mechanism for optimization. The model automatically extracts key features from complex demographic and carbon emission data, leveraging the attention mechanism to assign appropriate weights, thereby capturing the behavioral patterns and trends in energy consumption driven by population dynamics more effectively. By integrating multi-source data, including historical carbon emissions, population density, demographic trends, meteorological data, and economic indicators, experimental results demonstrate the model's outstanding performance across multiple datasets.

Estimated Time of Arrival (ETA) is a crucial task in the logistics and transportation industry, aiding businesses and individuals in optimizing time management and improving operational efficiency.... Show More

Estimated Time of Arrival (ETA) is a crucial task in the logistics and transportation industry, aiding businesses and individuals in optimizing time management and improving operational efficiency. This study proposes a novel Graph Recurrent Neural Network (GRNN) model that integrates external factor data. The model first employs a Multilayer Perceptron (MLP)-based external factor data embedding layer to categorize and combine influencing factors into a vector representation. A Graph Recurrent Neural Network, combining Long Short-Term Memory (LSTM) and GNN models, is then used to predict ETA based on historical data. The model undergoes both offline and online evaluation experiments. Specifically, the offline experiments demonstrate a 5.3% reduction in RMSE on the BikeNYC dataset and a 6.1% reduction on the DidiShenzhen dataset, compared to baseline models. Online evaluation using Baidu Maps data further validates the model's effectiveness in real-time scenarios. These results underscore the model's potential in improving ETA predictions for urban traffic systems.

At present, entrepreneurial ecosystem is becoming the frontier integration framework of entrepreneurship research. However, due to the lack of research results on the interaction and integration... Show More

At present, entrepreneurial ecosystem is becoming the frontier integration framework of entrepreneurship research. However, due to the lack of research results on the interaction and integration between different levels of the system, the strategic orientation of new ventures in the system is difficult to be consistent with the general trend reflected by the system. Based on this, combined with the existing research deficiencies, this paper focuses on the strategic orientation of new ventures in the entrepreneurial ecosystem, and proposes the concept of entrepreneurial ecological orientation to promote the mutual integration and isomorphism between new ventures and their surrounding entrepreneurial ecosystems. When a new venture is perfectly integrated into its entrepreneurial ecosystem, it will no longer be an isolated individual, but will become a co-owner of the system's interests. The state of integration and isomorphism with the entrepreneurial ecosystem will make the new venture no longer vulnerable, and the risk of entrepreneurial failure will be greatly reduced.

The paper focuses on how technology readiness affects the desire of bank workers in China to use artificial intelligence (AI) in the finance sector. It aims to show what factors are important and... Show More

The paper focuses on how technology readiness affects the desire of bank workers in China to use artificial intelligence (AI) in the finance sector. It aims to show what factors are important and how they relate to each other. The study used SMART PLS 4 to analyze the connections between these factors. Data was collected from 374 employees working at various levels of banking and financial sector organizations situation in the provinces of China. Trust in technology also plays a big role in this connection. For future studies, it would be good to look at how gender diversity policies and career paths change over time and to improve how we measure these things by including new factors like the negative sides of technology.

In digital advertising, accurately capturing consumer preferences and generating engaging, personalized content are essential for effective ad optimization. However, traditional methods often rely... Show More

In digital advertising, accurately capturing consumer preferences and generating engaging, personalized content are essential for effective ad optimization. However, traditional methods often rely on single-modal data or static models, limiting their adaptability to dynamic consumer behavior and complex, multi-dimensional preferences. To address these challenges, we propose a Multi-modal Adaptive Generative Adversarial Network for Ad Optimization and Response Prediction (MAGAN-ORP). MAGAN-ORP integrates multi-modal data—including text, image, and behavioral features—into a unified framework, enabling a comprehensive understanding of consumer preferences. The model includes an adaptive feedback mechanism that dynamically refines ad content based on real-time consumer interactions, ensuring relevancy in evolving environments. Additionally, a consumer response prediction module anticipates user engagement, allowing for proactive optimization of ad strategies.

The purpose of this research is to Explore the Association between Cultural Intelligence and Project Performance with the Mediation of Artificial Intelligence Abilities. Researchers have analyzed... Show More

The purpose of this research is to Explore the Association between Cultural Intelligence and Project Performance with the Mediation of Artificial Intelligence Abilities. Researchers have analyzed the data with SMART PLS 4.0 to investigate the connections among cultural intelligence, and artificial intelligence abilities, on project performance. We used a quantitative methodological design based on positivism to investigate the systematic impact of Cultural intelligence on project performance. A survey consisting of different positions performing project management activities, 374 project managers from various reputable construction industries was conducted. The findings emphasize the impact of cultural intelligence on project performance within the construction industry, implying that companies that manage to successfully infuse Cultural intelligence with artificial intelligence abilities while nurturing a culture based on artificial intelligence abilities can increase their project performance as well.

With the rapid growth of e-commerce and social media, analyzing consumer sentiment and predicting purchase intentions are vital for understanding behavior and optimizing marketing strategies. This... Show More

With the rapid growth of e-commerce and social media, analyzing consumer sentiment and predicting purchase intentions are vital for understanding behavior and optimizing marketing strategies. This study proposes an integrated model combining Bidirectional Encoder Representation from Transformers (BERT) and Deep Bidirectional Long Short-Term Memory Network (DBLSTM). BERT efficiently extracts semantic features from consumer reviews using self-attention mechanisms, while DBLSTM captures sequential dynamics, leveraging temporal dependencies for predicting purchase intentions. The model was evaluated on real-world datasets and compared against other deep learning models. Results showed that the BERT-DBLSTM model outperformed others in sentiment analysis accuracy and purchase intention prediction, demonstrating higher generalization and prediction accuracy. This approach provides enterprises with precise market insights, enabling improved marketing strategies, enhanced user satisfaction, and increased conversion rates.

This study systematically reviews the literature on the impact of technological innovations in e-commerce on consumer behavior, using the SPAR-4-SLR methodology and TCCM framework. It consolidates... Show More

This study systematically reviews the literature on the impact of technological innovations in e-commerce on consumer behavior, using the SPAR-4-SLR methodology and TCCM framework. It consolidates research across various technologies, including websites, social media, live streaming, AR/VR, and AI. The analysis reveals a growing interest in this field post-2017, with a focus on websites and social media, but highlights a research gap in emerging technologies. Key theoretical frameworks are identified, emphasizing the need for integration to comprehensively understand consumer behavior. The review maps out antecedents, mediators, moderators, and outcomes, stressing the importance of longitudinal studies and advanced analytics. This approach aims to bridge research gaps and suggest future directions, enhancing theoretical and practical understanding of e-commerce technological innovations, and contributing to a more dynamic and consumer-centric e-commerce ecosystem.

Financial markets are inherently complex and influenced by a variety of factors, making it challenging to predict trends and detect key events. Traditional models often struggle to integrate both... Show More

Financial markets are inherently complex and influenced by a variety of factors, making it challenging to predict trends and detect key events. Traditional models often struggle to integrate both structured, or numerical, and unstructured, or textual, data; additionally, they fail to capture temporal dependencies or the dynamic relationships between financial entities. To address this, the multidimensional integrated model for financial text mining and value analysis (MI-FinText), was proposed. MI-FinText integrated multi-task learning, temporal graph convolutional networks and dynamic knowledge graph construction. MI-FinText simultaneously performed sentiment analysis, event detection, and value prediction by learning shared representations across tasks and modeling time-dependent relationships between financial events. MI-FinText continuously updated a dynamic knowledge graph to reflect the evolving financial landscape, enabling real-time insights.

In the digital intelligence era, user participation in innovation is increasingly prevalent, underscoring the importance of user creativity in driving organizational change. Despite its relevance... Show More

In the digital intelligence era, user participation in innovation is increasingly prevalent, underscoring the importance of user creativity in driving organizational change. Despite its relevance, user potential creativity—a critical factor in digital product innovation—remains insufficiently defined and measured. To address this gap, this study refines the concept of user potential creativity and introduces a preference revelation approach using machine learning. Drawing on data from Thingiverse and proxy variables such as intrinsic motivation and lead userness, the research combines machine learning with ordinary least squares regression to examine the impact of user potential creativity on digital product innovation. The findings reveal that intrinsic motivation significantly enhances innovation performance. The study argues that managing user creativity requires attention to both user capabilities and motivations, and it proposes user management adaptive change as a new framework for organizational adaptation.

With the rapid rise of online recruitment, automating resume-job matching is crucial for hiring efficiency. However, current methods fail to capture complex semantic relationships, lack hierarchical... Show More

With the rapid rise of online recruitment, automating resume-job matching is crucial for hiring efficiency. However, current methods fail to capture complex semantic relationships, lack hierarchical representations, and overlook bidirectional compatibility. To address these issues, the authors propose a resume-job matching model integrating generative pretrained transformer 4 embeddings, hierarchical graph neural networks, and a bilateral matching algorithm. This approach effectively models semantic, structural, and bidirectional relationships. Experiments on public datasets show that the authors' model achieves an F1 score of 0.91, outperforming cosine similarity (0.70) and SuperGlue (0.88). Statistical analysis validates the improvements, while ablation studies highlight each module's role. Results show that the authors' approach enhances accuracy and reduces bias, providing a data-driven solution for fair and efficient recruitment.

With the development of e-commerce, the precision of accurately predicting user purchasing trends has been of great significance for personalized recommendation and marketing. However, the... Show More

With the development of e-commerce, the precision of accurately predicting user purchasing trends has been of great significance for personalized recommendation and marketing. However, the traditional purchasing behavior prediction methods are faced with the challenges of sparse data, long-term dependence modeling difficulties, and difficulty in effectively capturing the complex purchasing patterns and long-term behavior trends of users. In this paper, a user purchase trend analysis model combining self-supervised learning and convolutional neural network-long short-term memory architecture is proposed to solve the shortcomings of traditional methods in processing time series data and unlabeled data. Firstly, self-supervised learning is used to pre-train a large number of unlabeled user behavior data and extract potential features. The convolutional neural network-long short-term memory architecture is then used to combine local features and long-term dependencies to accurately predict user purchasing trends.

The rapid growth of large-scale information and the dynamic nature of user behaviors pose significant challenges for modern information retrieval systems, which often struggle to adapt to... Show More

The rapid growth of large-scale information and the dynamic nature of user behaviors pose significant challenges for modern information retrieval systems, which often struggle to adapt to non-stationary environments and fail to fully utilize multimodal data, leading to suboptimal performance. To address these issues, this study proposes the adaptive deep reinforcement learning (RL) framework for information retrieval and management, which combines RL, multimodal data fusion, and an adaptive update mechanism to dynamically adjust to evolving user preferences and document collections. The adaptive deep RL framework for information retrieval and management employs a RL-based policy network to optimize retrieval strategies, a multimodal encoder to integrate diverse data sources, and an adaptive mechanism to maintain robustness in dynamic scenarios.

Anomaly detection plays a critical role in fields such as finance, healthcare, and industrial monitoring, where identifying irregular patterns can prevent system failures and enhance security.... Show More

Anomaly detection plays a critical role in fields such as finance, healthcare, and industrial monitoring, where identifying irregular patterns can prevent system failures and enhance security. Traditional methods, however, often struggle with handling multimodal data and adapting to dynamic environments where data distributions change over time. Additionally, many models face difficulties in balancing false positive and false negative rates, particularly when dealing with imbalanced or noisy datasets. This paper proposes a novel approach, MDA-GAN (Multimodal Dynamic Autoencoder with Generative Adversarial Networks), to address these challenges. The method integrates deep autoencoders for feature extraction, utilizes GANs to generate high-quality normal samples, and incorporates dynamic threshold adjustment to enhance adaptability to evolving data patterns. By combining these components, MDA-GAN improves anomaly detection accuracy, robustness, and adaptability, particularly in complex multimodal data environments.

Although many studies have examined social status in offline contexts, the combined influence of digital metrics—such as the number of followers a person has—and traditional cues, such as expertise... Show More

Although many studies have examined social status in offline contexts, the combined influence of digital metrics—such as the number of followers a person has—and traditional cues, such as expertise on online status perceptions, remains underexplored. Drawing on the stereotype content model, this study investigated how expertise and number of followers jointly affect perceived social status, dominance, and warmth across both real-life and social media contexts. The findings show that high expertise and a large number of followers significantly boost perceptions of status and dominance yet have no meaningful impact on warmth. These results underscore the interplay between digital and traditional status cues and offer practical insights for personal branding, influencer marketing, and social media platform design.

This study investigates the intellectual core of mobile payment (MP) research through citation analysis, co-citation analysis, cluster analysis, and multidimensional scaling (MDS), based on 111... Show More

This study investigates the intellectual core of mobile payment (MP) research through citation analysis, co-citation analysis, cluster analysis, and multidimensional scaling (MDS), based on 111 highly cited articles published between January 1996 and December 2023 in the Web of Science (WoS) database. The analysis reveals 13 core knowledge clusters: (1) mobile money, (2) trust in mobile banking, (3) risk factors in digital banking, (4) service quality, (5) UTAUT-based decision frameworks, (6) beliefs, (7) IT adoption decision-making, (8) trust, (9) perceived value, (10) compatibility, (11) relative advantage, (12) social influence, and (13) intention to use and continued use of information systems. These clusters reflect the evolution of MP research from foundational theories to practical, user-oriented applications. By mapping key themes and identifying influential research directions, this study offers valuable insights for scholars and practitioners, contributing to a deeper understanding of the field and providing a structured basis for future research.

The exponential growth of e-commerce platforms has generated vast amounts of user behavior data, making it increasingly important to predict consumer preferences and spending patterns. Traditional... Show More

The exponential growth of e-commerce platforms has generated vast amounts of user behavior data, making it increasingly important to predict consumer preferences and spending patterns. Traditional recommendation systems often struggle with challenges such as data sparsity, the cold-start problem, and the inability to capture the dynamic nature of user behavior. These limitations hinder the accurate prediction of consumer actions, especially in evolving markets where user preferences change over time. To address these challenges, the authors propose deep behavioral and sentiment-aware personalized recommendation model, a novel approach that integrates dynamic user behavior modeling and sentiment analysis within a hybrid recommendation framework. The model leverages both collaborative filtering and content-based filtering, enhanced by deep learning techniques, to continuously adapt to evolving user preferences and emotional context, improving both recommendation relevance and consumer spending prediction.

Rapid technological advancements have transformed global education, bringing both opportunities and challenges, such as technostress. This study explores how educators' technological innovation... Show More

Rapid technological advancements have transformed global education, bringing both opportunities and challenges, such as technostress. This study explores how educators' technological innovation intention and emotional responses impact innovation performance in e-learning environments, focusing on emotional strain from information overload and technological complexity. Using information processing theory and data from 592 Chinese higher education educators, the study finds that emotional cognition and emotional management mediate the relationship between technological innovation intention and innovation performance. These results highlight the need to address educators' emotional factors for effective technological integration in education. The study contributes to the education and technology fields by extending information processing theory, showing how emotional dynamics influence innovation, and offering strategic recommendations for policymakers and leaders to reduce technostress, enhance innovation, and improve performance through emotional support and technology adoption strategies.

Many governments have adopted information intelligent technology (IIT) policies to enhance efficiency and responsiveness. This study examines key factors influencing IIT-based decision making in... Show More

Many governments have adopted information intelligent technology (IIT) policies to enhance efficiency and responsiveness. This study examines key factors influencing IIT-based decision making in Korean central and local governments. The research model includes perceived usefulness and risk of IIT, leaders' and peers' IIT, and social capital as moderators. Using quota sampling, regression analysis reveals that perceived usefulness, perceived risk, and leaders' IIT positively affect decision making. While bonding social capital has no moderating effect, bridging social capital enhances the impact of peers' IIT but weakens the effects of perceived risk and leaders' IIT. These findings highlight the role of personal perception, social influence, and social capital in IIT adoption, offering insights for future digital governance research.

The purpose of this research is to evaluate the effect of artificial intelligence (AI)-facilitated leadership on team member productivity. To realize this goal, a quantitative methodological design... Show More

The purpose of this research is to evaluate the effect of artificial intelligence (AI)-facilitated leadership on team member productivity. To realize this goal, a quantitative methodological design based on positivism was used to investigate the systematic impact of AI-empowered leadership practices on teamwork dynamics and performance determinants. A survey consisting of different cadres, i.e., 308 lower, middle, and top-level managers, as well as team members from different fast-moving consumer goods (FMCG) firms, was conducted. The outcome indicated a strong and positive connection between AI-backed leadership and team efficiency through a good big data culture. The findings emphasize the transformative impact of AI on leadership positions within the FMCG industry, implying that companies that manage to successfully infuse AI while nurturing a culture based on data can increase their competitiveness, as well. This information is crucial for FMCG firms aiming at utilizing AI and big data to improve team performance and ultimately achieve organizational success.

As artificial intelligence continues to advance, virtual streamers have emerged as a new trend in virtual live streaming commerce. However, limited research focuses on the specific characteristics... Show More

As artificial intelligence continues to advance, virtual streamers have emerged as a new trend in virtual live streaming commerce. However, limited research focuses on the specific characteristics virtual streamers should possess and how these characteristics engage consumers in this context. To address this gap, this study applies computers as social actors theory to examine how virtual streamer attributes (likeability, animacy, intelligence, and responsiveness) affect consumer engagement. Survey data collected from 382 Chinese virtual live streaming commerce consumers were analyzed using the partial least squares-structural equation modeling technique. The results reveal that likeability, intelligence, and responsiveness positively impact consumers' immersion, while animacy does not. The study also found moderating effects of virtual streamer-background congruence and virtual streamer-product congruence on these relationships. These findings contribute to virtual human literature and provide several actional suggestions for practitioners leveraging virtual streamers.

With the increasing complexity of enterprise systems and the rise in cyber threats, managing security risks while optimizing resources has become a significant challenge. Traditional models often... Show More

With the increasing complexity of enterprise systems and the rise in cyber threats, managing security risks while optimizing resources has become a significant challenge. Traditional models often address security and resource management in isolation, making it difficult to adapt to evolving threats and dynamic workloads. This paper proposes the deep learning-based dynamic security assessment and optimization model, which integrates dynamic security assessment, anomaly detection, multi-modal data fusion, security investment optimization, and cloud resource optimization into a unified framework. By leveraging deep learning techniques such as convolutional neural networks for feature extraction and recurrent neural networks for temporal anomaly detection, alongside reinforcement learning for resource optimization, the deep learning-based dynamic security assessment and optimization model provides real-time risk evaluation and adapts resource allocation based on system needs.

Social media has become a critical platform for brand management, enabling real-time consumer engagement, sentiment analysis, and influencer-driven marketing. However, existing models often focus on... Show More

Social media has become a critical platform for brand management, enabling real-time consumer engagement, sentiment analysis, and influencer-driven marketing. However, existing models often focus on isolated tasks, failing to capture evolving user interactions and sentiment shifts dynamically. To address this, we propose Hybrid Attention-based Personalized Deep Neural Network (HAP-DNN), a unified deep learning framework integrating hierarchical attention mechanisms, hybrid optimization, and personalized brand recommendations. HAP-DNN dynamically prioritizes engagement signals and sentiment cues while employing Moth-Flame Optimization (MFO) for adaptive hyperparameter tuning, improving training efficiency. A transformer-based recommendation module enhances brand suggestions by incorporating real-time user interactions. Extensive experiments demonstrate that HAP-DNN outperforms five state-of-the-art baselines, achieving a 17.5% lower MSE in engagement prediction, a 4.1% improvement in sentiment classification accuracy, and a 6.8% increase in brand loyalty prediction performance.

Personalized recommendation systems are crucial for improving user experience and business performance in e-commerce. However, existing models face two major challenges: (a) an imbalance between... Show More

Personalized recommendation systems are crucial for improving user experience and business performance in e-commerce. However, existing models face two major challenges: (a) an imbalance between short-term ranking accuracy and long-term engagement optimization, and (b) limited utilization of multi-modal information, resulting in suboptimal contextual understanding and poor cold-start performance. Traditional sequential models effectively capture short-term user preferences but fail to consider long-term engagement dynamics, while reinforcement learning (RL)-based approaches optimize engagement but suffer from high computational complexity and slow convergence. To address these challenges, we propose HDL-RecBERT, a hybrid recommendation framework that integrates transformer-based sequential modeling, RL for long-term optimization, adaptive multi-modal feature fusion, and contrastive self-supervised learning. The model employs self-attention mechanisms to model user behavior, RL to maximize cumulative user engagement, cross-modal attention to dynamically fuse multi-modal data, and contrastive learning to enhance cold-start recommendation performance. Extensive experiments on real-world e-commerce datasets show that HDL-RecBERT outperforms state-of-the-art baselines, achieving an 11.2% improvement in HR@10 and a 9.5% increase in NDCG@10, while RL improves cumulative reward (CR) by 10.2% and contrastive learning enhances cold-start recall by 9.3%. These results demonstrate HDL-RecBERT's ability to balance short-term and long-term optimization, improve recommendation diversity, and enhance adaptability in cold-start scenarios, making it a promising solution for next-generation recommendation systems.

In remote sensing, images are widely used in applications, such as land cover classification, urban monitoring, and disaster management, providing rich information about the Earth's surface.... Show More

In remote sensing, images are widely used in applications, such as land cover classification, urban monitoring, and disaster management, providing rich information about the Earth's surface. However, due to data heterogeneity and scarcity, different modalities of remote-sensing images often face challenges in classification tasks. The proposed deep learning model for remote-sensing image classification addresses these challenges through multimodal fusion. By combining a convolutional neural network, a generative adversarial network, and a graph convolutional network, the model is organized into three main components: data preprocessing and feature extraction, multimodal data generation and enhancement, and multimodal feature fusion and classification. Experimental results on the Hyperspectral-Light Detection and Ranging Houston2013 dataset and the Hyperspectral-Synthetic Aperture Radar Berlin dataset show that the proposed method significantly outperforms traditional methods and other deep learning models in classification performance, with better stability and robustness.

With the continuous improvement of international law, cross-border e-commerce platforms have been continuously developing and improving, becoming one of the main ways of international trade and... Show More

With the continuous improvement of international law, cross-border e-commerce platforms have been continuously developing and improving, becoming one of the main ways of international trade and economy. Online reviews substantially influence purchase decisions, yet their reliability is often eroded by manipulation, bias, and fraud. Prevailing methods depend mainly on sentiment analysis of review text, neglecting behavioral cues, probabilistic signals, and cross-platform variability. This study proposes the review credibility–trust integrated model—a comprehensive framework that unifies multi-dimensional credibility scoring, consumer trust formation, perceived risk estimation, and contextual factors (e.g., seller reputation, product category). The review credibility–trust integrated model synthesizes textual, behavioral, and probabilistic evidence to quantify review credibility and leverages structural equation modelling to disentangle the relationships among credibility, trust, risk perception, and purchase intention.

The widespread adoption of algorithm-driven short video platforms has transformed how users process content and engage in impulsive purchasing. However, current models often fail to capture the... Show More

The widespread adoption of algorithm-driven short video platforms has transformed how users process content and engage in impulsive purchasing. However, current models often fail to capture the multi-dimensional psychological mechanisms that underlie such behavior, particularly the dynamic interplay between content signals, emotional immersion, perceptual intuition, and algorithmic perception. To address these limitations, this study proposes stimulus-organism-response–elaboration likelihood model, an integrated behavioral model that fuses the stimulus–organism–response paradigm with dual-route elaboration (central and peripheral), flow experience, and algorithmic trust. The model differentiates between cognitive and behavioral impulsivity, incorporates post-purchase regret as a feedback mechanism, and models personalization and transparency as precursors to algorithmic attitude and trust.

In the context of data law compliance requirements, personalized recommendation systems have become integral to modern e-commerce platforms, yet most existing models rely solely on behavioral data... Show More

In the context of data law compliance requirements, personalized recommendation systems have become integral to modern e-commerce platforms, yet most existing models rely solely on behavioral data and overlook the affective and cognitive dimensions of user decision-making. This limitation leads to inadequate personalization, poor generalization in cold-start scenarios, and a lack of real-time adaptability under data law frameworks. To address these challenges, the paper proposes MTL-SA, a multitask learning framework that integrates behavioral signals, sentiment-aware representations, and reinforcement learning into a unified recommendation architecture. This study demonstrates that integrating affective and behavioral feedback through multitask architectures can significantly enhance the accuracy, robustness, and human alignment of personalized recommendation systems under data legal frameworks.

This study investigates the shifting role of digital platforms in enabling entrepreneurship in China's knowledge economy within the context of inclusivity, sustainability, and equity. Drawing on... Show More

This study investigates the shifting role of digital platforms in enabling entrepreneurship in China's knowledge economy within the context of inclusivity, sustainability, and equity. Drawing on qualitative and quantitative methods, including multiview augmented reality, visualization/augmented reality focus groups, and regression analysis, the study assesses urban and rural entrepreneurial dynamics, the sustainability of crowdfunding models over time, and resource allocation across socioeconomic and geographical contexts through digital ethnography and focused area analysis. Results reveal considerable differences between urban and rural entrepreneurs, with urban customers profiting from advanced infrastructure and analytics tools. Conversely, rural entrepreneurs are impeded by a lack of internet connectivity and digital literacy. This paper adds to the understanding of network effects and resource allocation among platform ecosystems. It emphasizes the importance of creating inclusive policies and platform innovations.

Carbon markets play an important role in combating climate change, and accurate carbon price forecasts and risk assessments are essential for effective policymaking and green finance. However... Show More

Carbon markets play an important role in combating climate change, and accurate carbon price forecasts and risk assessments are essential for effective policymaking and green finance. However, existing models face challenges in dealing with the complexity of carbon markets, especially in handling the integration of price data with structured and unstructured data such as policy documents. To overcome these problems, this paper proposes an innovative hybrid model, combining an autoencoder, bidirectional encoder representations from transformers (BERT) model, and Monte Carlos tree search (MCTS), called AutoBERT-MCTS. An autoencoder is used to extract market data features, BERT handles policy texts, and MCTS optimizes investment strategies in uncertain environments. The model effectively integrates multi-source data and optimizes the strategies, which is a powerful tool in the field of green finance. By enhancing carbon market prediction and risk management capabilities, this paper provides valuable insights for policy and investment decisions in climate change.

Enhancing user emotional experience is crucial for improving product appeal and user satisfaction. However, current systems often struggle to integrate explicit and implicit interactions... Show More

Enhancing user emotional experience is crucial for improving product appeal and user satisfaction. However, current systems often struggle to integrate explicit and implicit interactions, particularly in capturing subtle emotional signals and dynamically adjusting feedback based on user behavior. To address these challenges, XFormerLite, a lightweight transformer-based model, is proposed to optimize emotional experiences via explicit–implicit signal integration, temporal modeling, and adaptive feedback. By combining explicit signals (e.g., voice and gestures) with implicit signals (e.g., facial expressions and posture), XFormerLite improves emotional engagement and personalizes real-time feedback. Experimental results indicate that XFormerLite achieves an emotion recognition accuracy of 86.2% on the Emo-DB dataset and 84.5% on IEMOCAP, along with significant improvements in response speed and emotion onset delay (EOD). This model presents a promising solution for optimizing emotional experiences in human–product interactions.

This article presents an analysis that is based on the perspective of Marxist consumption theory. The exponential growth of e-commerce has heightened the problem of information overload, whereby... Show More

This article presents an analysis that is based on the perspective of Marxist consumption theory. The exponential growth of e-commerce has heightened the problem of information overload, whereby consumers face excessive and complex information in the form of product listings, descriptions, and reviews. This overload frequently leads to hesitation, decision delays, and decreased satisfaction. Existing models often treat overload as either a cognitive or an emotional issue and thus fails to account for their combined impact. To address this gap, the authors propose the Multilayered Information Overload and Consumer Decision Behavior Model (MIO-CD), which is based on Marxist consumption theory. This research offers a scalable, evidence-based framework for understanding consumer decisions and provides actionable insights for optimizing digital commerce experiences.

Timely and accurate access to financial data is crucial for empirical research in accounting and finance. However, current data collection processes are often manual, inconsistent, and difficult to... Show More

Timely and accurate access to financial data is crucial for empirical research in accounting and finance. However, current data collection processes are often manual, inconsistent, and difficult to scale. This study asks: How can large language models (LLMs) be effectively used to automate financial data collection? Using design science research methodology (DSRM), the author develops a modular architecture that integrates a real-time search API and auxiliary information processing into LLM workflows. The study applies the model to two tasks: extracting ESG report release dates and identifying customer firm tickers from COMPUSTAT. The system achieves 96% and 95% accuracy, respectively, comparable to human performance. This study advances LLM applications in accounting by providing a scalable, practical framework for automating financial data retrieval.

This research investigates the interplay between trust and control in the context of AI-enabled healthcare decision-making using a cybernetic approach to how clinicians adjust their trust toward... Show More

This research investigates the interplay between trust and control in the context of AI-enabled healthcare decision-making using a cybernetic approach to how clinicians adjust their trust toward AI-powered clinical decision support systems. It seeks to uncover the barriers to AI implementation within clinical practices regarding dependency and trust towards AI technologies. Forty-two clinicians who used an AI-powered clinical decision support system in radiology participated in an empirical mixed-method study. Participants interpreted x-ray images with different levels of AI control and transparency. Trust ratings, override behaviors, and decision accuracy were captured and analyzed with t-tests, ANOVA, and regression models. Trust dynamics were explored through semi-structured interviews. The findings reveal that when the AI adheres to clinical reasoning, its transparency boosts trust, yet it diminishes confidence when an error occurs.

With the increasing influence of social media in shaping consumer attitudes, modeling brand perception has become a critical task for computational marketing. Traditional approaches have... Show More

With the increasing influence of social media in shaping consumer attitudes, modeling brand perception has become a critical task for computational marketing. Traditional approaches have predominantly relied on textual sentiment analysis, which fails to capture the dynamic interplay of user interaction and platform-specific affordances. To address this limitation, the authors propose a novel Multi-Dimensional Perception Modeling (MPM) framework that jointly incorporates content semantics, social engagement signals, and platform-level features into an end-to-end neural architecture. MPM leverages BERT-based encoders for text, structured encodings for likes, comments, and influencer metrics, and visibility-aware metadata to represent platform context. These heterogeneous inputs are fused using attention mechanisms to predict both brand attitude categories and perception scores. Experiments on large-scale datasets from Weibo and Xiaohongshu demonstrate the effectiveness of the model.

With the rise of cryptocurrencies, illicit activities such as money laundering, fraud, and Ponzi schemes have gained attention. Traditional methods using graph neural networks (GNNs) to detect... Show More

With the rise of cryptocurrencies, illicit activities such as money laundering, fraud, and Ponzi schemes have gained attention. Traditional methods using graph neural networks (GNNs) to detect illicit transactions treat the entire transaction network as input, which works well on small networks but struggles with large-scale blockchain data. To address this limitation, the authors propose a neighborhood subgraph-based method that combines GCN and LSTM. The GCN captures information from neighboring nodes for each transaction, enhancing the understanding of the network structure, while the LSTM tracks the sequence and variations of fund flows. Experimental results show that by using 3-hop neighborhood subgraphs, the method outperforms other baseline models while requiring data from only an average of 80 nodes, thereby significantly improving efficiency compared to methods that process the entire transaction network.

This study focuses on the role of human capital development in promoting economic transformation and sustainability in five emerging economies: South Korea, China, Vietnam, India, and Brazil. Using... Show More

This study focuses on the role of human capital development in promoting economic transformation and sustainability in five emerging economies: South Korea, China, Vietnam, India, and Brazil. Using a mixed-methods approach, the research combines quantitative macroeconomic data with qualitative insights on how education, workforce training, innovation, and governance shape economic outcomes within various policies and institutional frameworks. South Korea and China have focused on developing knowledge- and technology-intensive industries and improving people's skills to meet this demand: STEM education, vocational training, or R&D investment. Vietnam presents a strong illustration of the power of targeted action, albeit with lower levels of tertiary enrollment and R&D investment. India and Brazil represent the challenges of actualizing human capital into economic fruits, thwarted by governance barriers and skills mismatches. This research elaborates on the pivotal link between human capital, institutional quality, global integration, and economic competitiveness.

Effective human resource management (HRM) is essential for optimizing enterprise decisions and enhancing employee satisfaction. However, traditional models rely on single-modal data and fail to... Show More

Effective human resource management (HRM) is essential for optimizing enterprise decisions and enhancing employee satisfaction. However, traditional models rely on single-modal data and fail to adapt dynamically to complex employee behaviors and emotions. To address these limitations, the authors propose HR-GENIE, a deep generative model integrating graph neural networks (GNNs), vision-language models (VLMs), and reinforcement learning (RL). GNNs analyze corporate social networks, VLMs extract emotional insights from multimodal data, and RL optimizes HR policies dynamically. Experiments show that HR-GENIE outperforms baseline models in MSE, RMSE, NDCG@10, employee attrition reduction, and NPS improvement, improving satisfaction prediction, social network analysis, and HR decision-making. Ablation studies confirm the contribution of each component. This study offers a data-driven HRM framework that enables enterprises to develop adaptive and employee-centric strategies, enhancing organizational stability.

This paper examines how AI adoption affects firms' labor investment efficiency, focusing on moderating the impact of globalization in GCC nations. Using firm-level panel data from 2015 to 2024, it... Show More

This paper examines how AI adoption affects firms' labor investment efficiency, focusing on moderating the impact of globalization in GCC nations. Using firm-level panel data from 2015 to 2024, it applies a System Generalized Method of Moments (GMM) to account for endogeneity and dynamic adjustments. Dynamic Capabilities Theory (DCT) suggests that AI adoption is a strategic resource for a firm to improve efficiency. Using four country-level globalization proxies: the Globalization Index (KOF), the Digital Adoption Index (DAI), the Global Competitiveness Index (GDI), and Total Factor Productivity (TFP), it found that globalization moderates AI adoption's impact on labor investment efficiency, supporting Globalization and Labor Market Theory, which states that globalization affects enterprises' resource optimization through global markets and technologies. It also shows that AI adoption improves labor investment efficiency significantly, and globalization's moderating effect strengthens this relationship, especially in more globally integrated countries.

Enterprises are increasingly venturing into financial investments beyond their core businesses, based on the non-parametric quantile model expanded by B-splines, to more truly reflect the nonlinear... Show More

Enterprises are increasingly venturing into financial investments beyond their core businesses, based on the non-parametric quantile model expanded by B-splines, to more truly reflect the nonlinear heterogeneous impact of organizational financialization and innovation investment. The findings reveal: (1) The financialization of enterprises has the convergence characteristics of the relationship between macro financial development and innovative investment in different economies. (2) With the improvement of the level of enterprise innovation investment (quintile point), the inflection point (or optimum point) of financialization gradually moves to the right, showing heterogeneous characteristics. (3) High-tech enterprises mainly exhibit the promoting effect of financialization, while non-high-tech enterprises mainly show the crowding-out effect. The financialization of Chinese enterprises should be commensurate with their own innovation level and macro financial development level, so as to provide a decision-making basis for improving the goal of innovation investment.

With the rapid development of mobile payments, understanding the dynamic relationship between consumer payment behavior and security perception is crucial for enhancing payment security and... Show More

With the rapid development of mobile payments, understanding the dynamic relationship between consumer payment behavior and security perception is crucial for enhancing payment security and optimizing user experience. This paper proposes the PaySenseNet model, which integrates Long Short-Term Memory (LSTM) networks, Deep Neural Networks (DNN), and a multi-task learning module, aiming to uncover the bidirectional interactive mechanism between payment behavior and security perception. The experimental results show that PaySenseNet outperforms traditional models across all datasets. On the PaySim dataset, the model achieved an accuracy of 88.6%, precision of 87.9%, MSE of 0.105, and a coefficient of determination R2 of 0.89; on the Credit Card Fraud Detection dataset, the accuracy was 87.3%, F1 score of 86.8%, MSE of 0.112, and R2 of 0.87; on the Mobile Payment User Behavior Data dataset, the accuracy was 88.1%, precision was 87.4%, MSE was 0.109, and R2 was 0.88.

This paper proposes the MT-DQN model, which integrates a Transformer, Temporal Graph Neural Network (TGNN), and Deep Q-Network (DQN) to address the challenges of predicting user behavior and... Show More

This paper proposes the MT-DQN model, which integrates a Transformer, Temporal Graph Neural Network (TGNN), and Deep Q-Network (DQN) to address the challenges of predicting user behavior and optimizing recommendation strategies in short-video environments. Experiments demonstrated that MT-DQN consistently outperforms traditional concatenated models, such as Concat-Modal, achieving an average F1-score improvement of 10.97% and an average NDCG@5 improvement of 8.3%. Compared to the classic reinforcement learning model Vanilla-DQN, MT-DQN reduces MSE by 34.8% and MAE by 26.5%. Nonetheless, the authors also recognize challenges in deploying MT-DQN in real-world scenarios, such as its computational cost and latency sensitivity during online inference, which will be addressed through future architectural optimization.

To tackle the limitations of traditional time-series models in capturing long-term dependencies, managing high computational costs, and adapting to volatile non-stationary financial data, this study... Show More

To tackle the limitations of traditional time-series models in capturing long-term dependencies, managing high computational costs, and adapting to volatile non-stationary financial data, this study introduces RiskMamba, a lightweight enterprise financial risk prediction model. It integrates Hierarchical Residual Multi-Scale Normalization (HRMS Norm) for stable gradient flow, Temporal Convolutional Networks (TCN) with dilated convolution to expand receptive fields for long-term dependency analysis, and the Mamba-2 Block to merge state-space modeling with parallel convolution for non-stationary sequence processing. Experiments on four datasets (Give Me Some Credit, S&P 500, Cryptocurrency Market, European Banking Stress Test) show RiskMamba outperforms baselines, achieving up to 15.7% lower MSE. By addressing non-stationarity and long-term dependency challenges, it offers a scalable framework for organizational risk management, enabling efficient edge computing deployment.

With the rapid development of digital economy, the scale and complexity of public data assets are increasing, and how to efficiently and accurately mine and utilize these data has become a key... Show More

With the rapid development of digital economy, the scale and complexity of public data assets are increasing, and how to efficiently and accurately mine and utilize these data has become a key issue. Existing methods are insufficient in dealing with multimodal data fusion and temporal feature capture, which is difficult to meet the needs of complex public decision-making. To this end, this paper proposes ReLSTM-Opt, a hybrid deep learning model that fuses ResNet and LSTM, to achieve automatic hyperparameter adjustment through Bayesian optimization, and to enhance the model's integrated learning ability for spatial and temporal features. Experimental results show that ReLSTM-Opt significantly outperforms traditional methods in classification accuracy and temporal prediction performance on multiple public datasets, demonstrating good generalization ability and stability. The model provides strong technical support for intelligent analysis and data-driven decision-making of public data assets and promotes the efficient utilization of public data resources in the digital economy.

This study explores how intelligent recommendation systems and platform governance mechanisms jointly influence crowdfunding success in the context of China's rapidly growing digital economy. Using... Show More

This study explores how intelligent recommendation systems and platform governance mechanisms jointly influence crowdfunding success in the context of China's rapidly growing digital economy. Using 300 structured survey responses and behavioral data from 5,000 real-world crowdfunding projects, the study investigates the critical roles of recommendation transparency, perceived platform control, and user learning capability in shaping investor trust and engagement. Adopting a mixed-method empirical framework, the findings reveal that enhancing algorithmic transparency and platform governance significantly fosters trust, which in turn improves crowdfunding outcomes. These insights offer actionable implications for policymakers and platform designers seeking to strengthen financial inclusion and support SME financing in emerging digital markets. The study contributes to the evolving discourse on digital platform governance by providing practical strategies for improving trust-building mechanisms in FinTech ecosystems.

In the Industry 5.0 era, human–machine collaboration, personalized intelligence, and sustainability are transforming HR and supply chain management. Grounded in the resource-based view and dynamic... Show More

In the Industry 5.0 era, human–machine collaboration, personalized intelligence, and sustainability are transforming HR and supply chain management. Grounded in the resource-based view and dynamic capabilities theory, this study examines how digital HRs (DHRs) influence supply chain agility (SCA) and innovation (SCI), with digital capabilities (DCs) as a mediator and organizational flexibility (OF) as a moderator. Survey data from 719 Chinese manufacturing firms show that DHRs positively impact SCA and SCI through enhanced DCs, which partially mediate the effect. While OF does not affect the DHR–SCA link, it negatively moderates the DHR–SCI relationship. This research advances cross-disciplinary understanding of digital HRs in supply chains and offers insights for firms optimizing resources amid intelligent transformation.

Urban traffic congestion remains a persistent challenge in smart cities, where complex spatiotemporal dependencies and limited interpretability hinder effective prediction and control. In... Show More

Urban traffic congestion remains a persistent challenge in smart cities, where complex spatiotemporal dependencies and limited interpretability hinder effective prediction and control. In law-governed road systems, effective prediction and intervention further depend on the accurate, automatic recognition of traffic laws (e.g., speed limits, lane-use rules, turn restrictions, signal timing plans), which provides the legal-operational constraints for intelligent control. Traditional models often fail to capture evolving causal relationships across road networks and lack the capability to simulate the impact of traffic interventions. To address these limitations, we propose DAST-CI, a Dual-Attention Spatiotemporal Causal Inference network that integrates dynamic causal graph learning, multi-modal feature fusion, and interpretable intervention reasoning.

Personalized advertising music is a cornerstone of digital marketing, yet existing strategies often rely on static segmentation or content heuristics that fail to scale with user complexity. The... Show More

Personalized advertising music is a cornerstone of digital marketing, yet existing strategies often rely on static segmentation or content heuristics that fail to scale with user complexity. The emergence of generative AI introduces new opportunities, but integrating it into a complete targeting pipeline remains a critical challenge. To address this, the authors propose 3P-IMF (Persona–Prompt–Performance Integrated Model Framework), a generative architecture that unifies behavior-driven persona segmentation, prompt-based content synthesis, and outcome-driven feedback loops. The framework uses Gaussian Mixture Models for soft clustering of user behaviors, structured prompt design conditioned on persona attributes, and iterative A/B/n testing to refine ad variants. To evaluate effectiveness, they compare 3P-IMF against five representative baselines across two large-scale datasets (Criteo, Avazu). Results show that 3P-IMF improves CTR by up to 52.9% and doubles ROI compared to traditional IMC and STP models.

Intelligent manufacturing systems are increasingly required to optimize both operational efficiency and supply chain resilience under uncertain and dynamic industrial conditions. However, current... Show More

Intelligent manufacturing systems are increasingly required to optimize both operational efficiency and supply chain resilience under uncertain and dynamic industrial conditions. However, current AI-driven methods often focus on isolated objectives, lacking integration across energy-aware scheduling, real-time disruption forecasting, and decentralized collaboration. To address these challenges, the authors propose a unified framework—RDAF-NSM (Resilient Dual-AI Fusion for Networked Smart Manufacturing)—which combines reinforcement learning, federated time-series modeling, semantic blockchain, and multi-objective evolutionary optimization. Specifically, the architecture integrates a Dual-AI Fusion Engine (DAF) for adaptive scheduling, a Resilience-Aware Supply Controller (RSC) for proactive anomaly management, a Knowledge-Augmented Blockchain Layer (KBCO) for secure and transparent inter-factory communication, and an Energy-Adaptive Decision Layer (EADL) for optimizing energy-resource tradeoffs.