Environmental Non-Governmental Organizations (ENGO) play a crucial role in advancing environmental protection by using threats and boycotts to force enterprises to produce sustainably. ENGO interact with enterprises based on different motivations and benefit from psychological gains, media publicity and the reduction of negative environmental externalities, but ENGO don't always victory and achieve their objectives. We use a game-theoretic model to examine how different types of ENGO motivation affect strategic interactions between an ENGO and a firm, and use a contest process to elucidate the boycott of ENGO to the firm. We find that, the result of the interaction is solely contingent upon their relative status, and the contest is most vigorous when they have the same status. The strategy of ENGO depends on both the primary motivation behind the engagement and the specifics of the target firm,and we categorize engagement actions as Radical engagement,Self-improvement engagement and Self-transcendence engagement base on the different engagement strategy. Our findings contribute to strategic management research by developing new insights about the interaction between ENGO and firm and the categorization of engagement actions,and indicate that the government need to regulate ENGO' s action sometimes.
In global financial governance, interdependence is mainly manifested in the increasing interconnection between international financial organizations, the broadening of the agenda of financial globalization, and the rise of financial development issues. Based on the convolutional neural network, this paper designs and implements a financial risk control model that can be applied to the actual production environment. The entire model has the specificity of dual input and multi-modality and can use international economic law and international relations as an effective input source to perform sufficient feature extraction so that the entire financial risk control model is based on cross-perspective prediction results. Experiments show that the data processing method in this paper can significantly improve the evaluation indicators of multiple models by at least 2
In this paper, I construct a macroeconomic agent-based model to explore its potential as a tool to support macroeconomic policy decisions. Compared with existing models, I include several channels of the transmission mechanism in the model such as interest rate channel and expectation channel. I then conduct a series of experiments to understand the responses of key macroeconomic variables to major macroeconomic shocks. I find that the effects of the demand shock and the monetary policy shock are broadly consistent with those of standard models. On the other hand, supply shock generates additional effects that are not typical for other classes of models.
E-commerce growth has increased rapidly in recent years, and several people utilize this popular channel to purchase services and products through Internet resources. Shopping sites have become very important for customers to buy the best products, and sales have increased for this resource. Furthermore, people purchasing through internet resources face several problems and a lot of confusion due to the massive number of products. They find it very difficult to choose their favourite product. In the present market, various popular traditional algorithms named Collaborative Filtering (CF), Planned Behavioral Theory (PBT), Markov Hidden Model (MHM), Traditional Machine Learning (TMC) and Analysis of Component (AoC) were used in the E-commerce sites for the users to purchase and choose the products in a customized manner with high service and more loyalty. In the traditional methods, Customers face several difficulties in determining the statistical probability of the product due to bulk data information during shopping. In this research, the heuristic computational method (FHCM) has been integrated with e-commerce sites, which helps optimize product search during the purchase and customer authentication process in an effective manner. This proposed method has been experimentally analyzed at lab scale testbed software and found to be more helpful in solving the problems in data sparse to identify the best product on the site for the customers.
Voting is a core element of social choice theory and a subarea of computational social choice. The goal is to rank (or rate) candidates according to voters’ preferences, and to eventually select the winner(s) of the election. In multiwinner voting, the same applies to sets of candidates and winning committee(s). The intuition is therefore quite similar to centrality in networks, where the goal is to rate (or rank) nodes—or groups of nodes—according to their structural positions. We establish correspondences between these two research fields by deriving preference rankings from network relations, and adapting single- and multiwinner voting rules to identify the most central (groups of) nodes. The transfer of reasonable and desirable properties and axioms from social choice theory to network science opens up the possibility to study novel aspects of centrality, and leads to the definition of voting-based centrality measures.
This research undertakes the study on how digitalization has been affecting labour market trends and social setups. As the digital economy expands, it tends to alter employment dynamics, skill formations, and the distribution of income among different sectors. The research investigates four primary hypotheses: (H1) The digitalization process creates more jobs than it destroys; (H2) Digitalization increases labour income shares through skill upgrading; (H3) Digitalization increases labour income share through the expansion of the private sector; and (H4) Digitalization worsens income inequalities, which leads to a reduction in the overall labour revenue stake. Through the exploration of the socio-economic effects of digitalization, this paper attempts to determine how technology-based changes affect labour market structures, employment possibilities, and wealth distribution. The study finds that, while digitalization provides ample jobs with increased wages, it also increases income imbalance between highly skilled and low-skilled workers. The findings therefore underscore the urgency of policy interventions addressing the emerging inequality to ensure a fair sharing of digitalization benefits.
Mediatization of music performance on short-video platforms highlights how digital technologies and platform-specific features shape the creation, presentation, and consumption of music. These platforms emphasize algorithmic visibility, audience interaction, and the construction of new performer identities. This study investigates how Chengdu musicians engage with such platforms to build identity and connect with audiences, while also examining differences in usage shaped by platform characteristics. Employing a mixed-methods approach, the research combines survey data from 674 respondents with thematic analyses of musician interviews. Quantitative analysis using Pearson’s and Spearman’s correlations revealed weak but significant associations between user engagement and entertainment value, as well as between user motivation and performer identity. Basic statistical results showed demographic differences—such as age, gender, occupation, and musical background—in shaping short-video activities. The qualitative findings demonstrated how musicians adjust performance aesthetics to align with algorithms, balance authenticity with visibility, and embed cultural identity in their content. Audiences valued entertaining, authentic, and culturally relevant performances. While musicians acknowledged artistic limitations imposed by platform dynamics, they also recognized new opportunities for creative expression. Overall, the study underscores the transformative impact of digital mediatization and platformization on music performance and artist-audience relationships.
I use a computational simulation model to test and develop theory on Duncan’s conjectures regarding the degree of uncertainty experienced by managers under varying task complexity and environmental turbulence. Thereby, Duncan’s conjectures are formally validated, with some qualifications. The simulation results further suggest that uncertainty can be lowered significantly by switching from a maximizing approach to a satisficing approach, trading off a modest extent of the probability of organizational success. This research has the potential to pave the way for resolving the deadlock between perceptual and objective measures of uncertainty—by placing environmental uncertainty in a logical framework so that it is operationalized more effectively. It enables giving credit to managers where due, by considering the level of uncertainty overcome in arriving at an organizational outcome. The unique contribution of the theory lies in involving multiple actors, considerations of limits to knowledge, and further consideration of multiple preferences.
This paper will look at one of the applications of an Artificial Intelligence (AI) in e-commerce websites, a specialization on user AI-driven systems such as product recommendations systems, chatbots, voice assistants, and AI-enhanced search filters. A systematic online survey was conducted by 5,000 online consumers who visited AI applications on e-commerce websites such as Amazon, Flipkart, and JD.com. The study had measured the user attitude towards personalization, convenience, trust and satisfaction towards the AI features. Results show that the sample is divided into large parts, namely 2635 (37.5) and 1825 (32.1) years old, with the former being digitally active and having a high level of use of AI technologies. The breakdown of incomes reflects that 33.4
Pink slime journalism is a practice where news outlets publish low-quality or inflammatory partisan articles, claiming to be local news networks. This paper examines the spread of pink slime sites on Facebook using public posts from Pages and Groups via the CrowdTangle API. We evaluate the trends of sharing pink slime sites on Facebook and patterns regarding the advertisements purchased by the parent organizations of the pink slime news networks using data from the Facebook Ad Library. Our over time data analysis discovers that while the number of pink slime posts on Facebook pages have decreased over the years, advertising dollars have increased, particularly in swing states. Furthermore, through analyzing the demographic data of the Facebook ad impressions, we see parent organizations like American Independent and Courier targeting a younger and more female population than Metric Media. Using Pearson correlation, we find that the increase in advertising dollars is associated with an increase in Facebook group posts. Further, the advertising expenditure increases during election years, but contentious topics are still discussed during non-election years. By illustrating geographic patterns and themes from US election years of 2020, 2022, and 2024, this research offers insights into emerging ‘local’ journalism tactics, and provides predictions for future US Presidential Elections.
Human Resource Management and effective people management are essential for evaluating and improving employee performance to meet organizational goals. However, current studies lack a thorough analysis of how employee well-being, such as mental health and work-life balance affects performance. Integrating AI, ML, and ethereum blockchain technologies can enhance the accuracy of these evaluations. So, this paper proposed a employee performance prediction and management system using 2(LS)M based approach. Initially, data collected from performance prediction dataset is pre-processed, and time series is analyzed using HARR-BIMA. Then, features are extracted from both pre-processed and time series analyzed data, followed by feature selection utilizing DACOA. Next, 2(LS)M predicts employee performance, and deviation is analyzed using the Z-score. If the deviation is high, employee mental health data is collected and pre-processed, and features are extracted. Pearson Correlation Coefficient is used to analyze the correlation between extracted mental health and performance prediction features. Optimal features from the extracted mental health data are then selected using DACOA, and performance is managed using 2(LS)M based on the correlation and deviation analysis, and optimal features. Meanwhile, all the information from both pre-processed mental health and performance data are protected using KWTCHT and securely stored on the ethereum blockchain. In experimental analysis, the proposed 2(LS)M achieved an accuracy of 98.65
The concept of multimodal human–computer interaction (HCI) and affective computing is now an innovative way of enhancing the analysis and reading of the public opinion in governance. Because of the limitations such as social desirability, anonymity, and a limited area to unearth the hidden sentiments, conventional methods such as surveys, polls, and town halls are often not sufficient to gather the richness of human feelings. These conventional means are often one-sided concerning the behavioral and emotional elements expressed through tone, facial expression, and other means, and instead focus only on the text delivery. It leads to poor or sometimes mistaken interpretations of the public opinion as important emotional data shaping the attitude and belief in the government are not explored completely. One of the solutions to these constraints is to introduce a multimodal framework that integrates advanced artificial intelligence methods with emotional computing. Vision Transformer -Convolutional Neural Network (ViT-CNN) is used to analyze facial expression and determine emotions by looking at the visual data, and Bidirectional Encoder Representations with Transformers (BERT) is used to interpret textual input and extract the context and emotionally loaded features. The framework would also help in providing a more in-depth understanding of the thoughts and emotions of the population with the inclusion of both the text and visual modalities. Moreover, it also incorporates a response generating mechanism powered by Flan-T5 to create sympathetic and environmentally aware feedback, which improves interactivity between citizens and the government and their trust. The results indicate that the proposed framework achieves an overall accuracy of 95.20
This paper discusses the creation of realistic, dynamic, and controllable synthetic social media data to support instruction on evaluating social-cybersecurity maneuvers in social media. We propose an agent-based simulation called SynX that takes as input the scenario templates created by Netanomics’ AI-Enabled Scenario Orchestration and Planning (AESOP) tool and outputs an X/Twitter API v1 message corpus by leveraging a large language model (LLM). We conduct an experiment on LLM prompting and evaluate the output of SynX using network metrics and the BEND framework.
In the group decision-making process in a social network, decision-makers interact within specific social relationships. Compared with the social relations, which is considered to have a significant influence on decision makers in previous studies, confidence, a personal trait, will also have a potential impact on both individual and collective interactions. Thus, this paper investigates the consensus model of social network group decision-making under the influence of confidence. First, given the widespread application of social platforms in facilitating group opinion interactions, this paper constructs an online social meta-network and integrates social network analysis with data mining techniques. By developing a confidence efficacy metric, it objectively quantifies the actual impact of decision-makers’ confidence on the decision-making process through the analysis of their online interactive behaviors. Second, the offline network structure efficacy is combined to construct a two-tier efficacy impact matrix for forming decision-makers’ weights. Then, Considering the positive role of confidence in opinion coordination, an opinion-sharing network is constructed based on “sharing willingness” (confidence efficacy) and “sharing channels” (network structure). This network identifies effective dissemination paths and utilizes opinion-sharing chains for precise feedback, thereby enhancing the efficiency and quality of consensus attainment. Finally, the proposed consensus reaching mechanism is applied to a community resident group decision-making case to verify the rationality and effectiveness of the proposed method. The case application results show that the decision-making model considering the influence of confidence significantly improves the efficiency and consensus quality of group decision-making, verifying the key role of the feedback mechanism driven by the opinion sharing chain in promoting information circulation and accelerating consensus formation.
Public health crises challenge information systems with rampant misinformation. This can trigger a “sensemaking collapse,” undermining organizational response and public trust. Traditional fake news detection methods, often decontextualized, are inadequate for these dynamic, knowledge-intensive events. This paper introduces EKAN (External Knowledge-Augmented Attention Neural Network), a novel framework grounded in situated cognition. EKAN functions as a computational augmentation tool. It integrates dynamic external knowledge from authoritative sources with internal content features via an attention mechanism. This process creates a “cognitive anchor” that enables context-aware veracity assessment. Validated on the Ohio Train Derailment dataset, EKAN significantly outperforms static baselines. This highlights the criticality of dynamic knowledge contextualization for robust countermeasures. It also offers a pathway toward designing more epistemically resilient crisis information systems that can support effective sensemaking amidst informational uncertainty.
This study examines the rhetorical strategies (ethos, pathos, logos) in face-to-face, written, and digital communication in South China. It explores how cultural and contextual factors influence these strategies and employs a mixed-methods approach to analyze rhetorical patterns. Data were collected from 200 participants through semi-structured interviews, written narratives, and digital communications. The findings indicate that Digital communication primarily appeals to emotions (pathos) through platforms such as social media and online messaging, while face-to-face and written communication prioritize credibility (ethos) and logic (logos). This research highlights the role of communication mediums and cultural contexts in shaping persuasive communication, providing new insights for the study of digital rhetoric and computational communication.
The growth of social networks has transformed social interactions, communication dynamics, and information dissemination. These social platforms have created complex systems in which user behavior plays an effective role in the information dissemination process. This paper introduces a novel CLAIPRD (Candidate, Latent, Active, Influential, Problematic, Recovered, and Detached) dissemination model, which incorporates three behavioral dimensions, peer influence (popularity-seeking), compulsive engagement (problematic use), and awareness (self-regulation). This seven-state framework captures the multifaceted roles of users in the dissemination process by explicitly modeling their behavioral transitions. A system of differential equations is formulated to describe the transitions among these states, enabling both qualitative and quantitative analyses. The equilibrium points and their stability properties are analyzed to determine the conditions under which information dissemination either stabilizes or persists. Additionally, the basic reproduction number is derived to quantify the potential for large-scale dissemination across the network, serving as a central element of the comprehensive dynamical analysis. The proposed model enhances understanding of the interplay between user behavior and information diffusion, offering a robust framework for analyzing and managing information flow in social networks. Numerical simulations are performed to validate the theoretical results, thereby establishing a solid foundation that deepens understanding of the complex dynamics underlying information diffusion in social networks.
Graduate education expansion in China brings quality management challenges, yet existing research predominantly employs static frameworks analyzing bilateral relationships, neglecting dynamic multi-stakeholder interactions and evolutionary mechanisms. We develop a tripartite evolutionary game model incorporating supervisors, students, and college administrations to examine how strategic choices evolve toward stable equilibria under bounded rationality. Through replicator dynamics analysis, Jacobian stability assessment, and numerical simulation, we identify five equilibrium scenarios including mutual disengagement, unilateral free-riding, and optimal cooperation. Results show that high-quality equilibria emerge when complementarity conditions are satisfied where net returns to serious effort exceed free-riding benefits for both supervisors and students, while administrative gains from strict oversight justify intervention costs. Sensitivity analysis reveals equilibrium structures depend on inequality constraint satisfaction rather than precise parameter values. Stochastic extensions demonstrate systems exhibit persistent fluctuations around deterministic states due to environmental uncertainties. Findings suggest sustainable quality improvement requires coordinated restructuring of incentive systems through reward calibration, cost reduction, and complementarity enhancement rather than intensified monitoring alone.
This study investigates whether and how hybrid organizations continuously achieve their dual objectives. Given the paradoxical pursuit of these goals within such organizations, we argue that hybridity is not a fixed concept; rather, its level fluctuates over time. From this dynamic perspective, we examine how hybridity can be temporally reconstructed. Specifically, since hybridity is shaped by two conflicting institutional logics, we explore how hybrid organizations manage this institutional complexity. Furthermore, we suggest that external incentive mechanisms can support these organizations in reconciling their inherent complexities. To analyze organizational responses to these mechanisms, we develop an agent-based model. This approach effectively addresses the diversity in hybrid organizations’ behaviors under dual-purpose settings. Through a qualitative investigation of social enterprises, we detail how these hybrid organizations handle their dual logics and respond to an externally influenced incentive system, illustrating how hybridity can be both enabled and limited by external incentives.
