Credit classification, which aims to characterize user profiles and accurately predict default risks, is critical for financial risk management. Although centralized models trained on directly aggregated multi-source data can achieve strong performance, they often disregard data privacy, class imbalance, and model interpretability. To address these gaps, we propose an example-dependent cost-sensitive vertical TabNet (ECS-VTabNet), a novel framework for privacy-preserving collaborative credit classification. The proposed framework integrates vertical federated learning with an attentive interpretable model (i.e., TabNet) and an example-dependent cost-sensitive (ECS) loss function. Our approach enables multiple institutions to collaboratively build a highly accurate credit scoring model without sharing raw data. The ECS mechanism aligns the model with business objectives by minimizing financial losses, and TabNet's inherent architecture provides transparent feature importance. Experimental results on two public credit datasets show that ECS-VTabNet outperforms existing federated and centralized methods in managing imbalanced data and providing interpretable predictions while avoiding raw data exchange under the honest-but-curious assumption. This work provides a practical solution for financial institutions seeking to build robust, compliant, and collaborative risk assessment models without centralizing sensitive client data.

PurposeDespite growing research on blockchain value, little is known about the role of CEOs in shaping these outcomes. This study examines whether CEO external directorships influence market reactions to blockchain investment announcements and how adoption strategy and application type condition this relationship through interaction effect.Design/methodology/approachWe conduct an event study of 200 blockchain-related announcements by 138 U.S. public firms and use regression analyses to assess the effects of CEO external directorships, adoption strategy and application type.FindingsCEO external directorships positively affect market responses to blockchain investments, with stronger effects for collaborative initiatives and financial applications. The results highlight the contingent value of CEO external social capital in complex IT investment decisions.Practical implicationsCEO external board ties can be a strategic asset in navigating regulatory, collaborative and technological complexities in blockchain projects. Boards and investors may consider executive network capital when evaluating leadership and strategic IT decisions.Originality/valueThis study offers one of the first pieces of empirical evidence on how CEO external directorships influence market reactions to blockchain investments. More importantly, it extends social capital, signaling and governance theories by showing that in blockchain investment contexts - characterized by heightened inter-organizational coordination, regulatory scrutiny and interpretive uncertainty - CEO external directorships function as context-activated social capital, whose value depends on adoption strategy and application type rather than being universally beneficial.

The dynamic pricing mechanisms of firms and search techniques for historical prices have spurred strategic behaviors and price expectations among customers. These customers predict future markdowns and delay their purchases based on price expectations (i.e., reference prices), where joint pricing and ordering decisions should be considered carefully to counteract or soften the negative impact of customers’ strategic behaviors and their reference prices on the seller’s profitability. Moreover, relevant literature neglects general and important features in the actual market (e.g., nonzero price thresholds and randomness in the formation of reference prices). In this study, we establish a Markov game between the retailer and strategic customers over a multi-period horizon, and set the reference price as a three-regime linear piecewise model with loss and gain thresholds. Then, we propose an adaptive multi-agent deep deterministic policy gradient (MA2DDPG) algorithm containing the experience replay buffer separation mechanism and delayed actor updates. Experimental results with synthetic and real-world datasets reveal that our algorithm converges to optimal policies in terms of convergence and rationality, and that it vastly outperforms the benchmark algorithms. Moreover, valuable managerial insights are obtained through an extensive numerical analysis for practitioners, particularly when they must consider the complicated characteristics of heterogeneous customer populations, such as disappointment behaviors, price thresholds, and strategic customer proportions. Our results pave the way for future research aimed at using multi-agent reinforcement learning to improve operations management with behavioral factors.

Anthropomorphism-defined as the attribution of humanlike characteristics such as identity, verbal and nonverbal behaviors, and other traits to nonhuman agents-plays a pivotal role in alleviating psychological discomfort in human-computer interactions. Although prior research in information systems (IS) has examined anthropomorphism, its development and application within IS and information technology (IT) contexts remain fragmented and underexplored. This study addresses the lack of a cohesive understanding of anthropomorphism in IT/IS environments, where systems are designed, deployed, and used. Drawing on the stimulus-organism-response (S-O-R) framework, we systematically review 195 studies published in leading IS journals, synthesizing existing insights and identifying key gaps. Our findings deepen the theoretical understanding of anthropomorphism in IT/IS and offer actionable guidance for designing humanlike features in digital systems. We conclude by presenting a comprehensive research agenda that outlines critical directions for advancing both theoretical and practical discourse on anthropomorphism in the IT/IS domain.

This study employs a Design Science Research approach to propose a foundational information system design theory tailored for Generative Artificial Intelligence (GenAI) applications in the fashion design process. It delineates meta-requirements and design principles that address both the transformative potential of GenAI, and the unique challenges faced by the fashion sector. To validate the practicality of the proposed design theory, a prototype system was developed and evaluated with feedback from 30 experienced fashion practitioners, confirming its feasibility and effectiveness. Insights from case studies conducted with two Hong Kong-based fashion companies further highlight the benefits and challenges of integrating GenAI into fashion design. While GenAI demonstrates promise in enhancing communication, accelerating design processes, and improving customer engagement and satisfaction, key challenges remain, including the need for high-quality datasets, significant computational resources, and ethical considerations related to AI-generated designs. The design principles derived from this study provide a structured guideline for system designers, offering a practical framework for developing GenAI systems that cater to the specific needs of the fashion industry. By contributing both theoretical and practical insights, this study advances understanding of how GenAI can drive innovation in fashion design and lays a foundation for future research in this domain.

This study draws upon the principal-agent theory to investigate the relationship between employee- related social performance and information security. This exploration encompasses both positive and negative dimensions of such performance: employee-related socially responsible activities (employee-related CSR) and employee-related socially irresponsible activities (employee-related CSiR). We employed a multistudy approach. First, we analyzed an eight-year sample of publicly listed firms, revealing a negative association between firms' engagement in employee-related CSR and information security risks, while their involvement in employee-related CSiR is positively linked to such risks. Our exploratory analysis uncovered additional intriguing findings, demonstrating that the uniqueness of employee-related social performance can amplify its impact on security. In a subsequent study, we conducted a scenario-based experiment to provide empirical evidence for our proposed principal-agent-based theory.

Background Federated Learning (FL) offers a privacy-preserving solution for multi-party data collaboration in smart healthcare. However, the data heterogeneity among hospitals and among patients often results in suboptimal performance for some hospitals when applying a global FL model. Current clustering-based FL methods struggle to adapt to complex and diverse data distributions, negatively impacting model performance. Methods We propose a novel framework, Federated Gaussian Mixture Clustering (FedGMC), which leverages Gaussian Mixture Clustering to train personalized FL models. FedGMC determines the optimal number of clusters prior to the FL process, reducing the time and computational cost associated with traversing multiple clustering configurations in existing approaches. Results The FedGMC framework was evaluated using real-world eICU datasets with various classifiers and performance metrics. Experimental results show that FedGMC outperforms other baseline methods in terms of the overall performance of combining two classifiers and two performance metrics. Moreover, it mitigates the risk of performance degraded for participating hospitals following FL. Conclusions The FedGMC framework effectively addresses clinical heterogeneity, enhancing predictive performance and ensuring fairness among participating medical institutions. These improvements increase the willingness of data owners to engage in the collaboration FL initiatives.

This study examines the role of dynamic capabilities (DCs) in the adoption of artificial intelligence (AI) systems in China’s telecommunications industry and their impact on firms’ operational performance. Drawing on DCs theory and the technology–organization–environment (TOE) framework, this study investigates how TOE factors influence DCs for AI system adoption, which subsequently impacts operational performance. It also examines the mediating role of DCs in the relationship between TOE factors and operational performance. Despite numerous studies on AI system adoption, the underlying mechanism of DCs requires further investigation. Survey data were collected from 205 senior executives of telecommunications firms in China. Analysis of the data reveals that TOE factors positively influence DCs for AI system adoption, which in turn have a positive effect on operational performance. Additionally, competitive pressure positively moderates the relationship between technological factors (i.e., technology compatibility and expected benefits) and operational performance. These findings provide managerial and theoretical insights into AI system adoption.

While international business studies suggest that multinationals conduct extensive market research and innovation tailored to bottom-of-the-pyramid (BOP) customer demand, the inherent challenges of the BOP market remain overlooked. This study highlights the role of product imitation in tackling the challenges of cost reduction and awareness improvement. The imitation literature focuses on the role of institutional environments in preventing product imitation in emerging markets, while little attention is paid to how customers—especially BOP customers, the dominant arbiters of value in emerging economies—motivate firms to engage in product imitation. Drawing on institutional theory and the demand-side view, this study examines the relationship between BOP orientation—the orientation toward meeting the unique demands of BOP customers—and product imitation. Using survey data from 334 Chinese manufacturing firms, this study finds that BOP orientation has a positive relationship with product imitation. Legal incompleteness and demand uncertainty strengthen this relationship, while demand heterogeneity weakens it. This study contributes to the literature by identifying BOP orientation as an important demand-side antecedent of product imitation and highlighting how these effects are contingent on institutional and demand factors.

The complex relationship between product or service attribute performance and customer satisfaction in online review environments has been widely discussed. However, attribute configuration for enhancing customer satisfaction under holistic thinking, particularly considering temporal interactions among attributes, remains underexplored. To address this gap, this study develops a dynamic configurational framework based on complexity and three-factor theories to deconstruct attribute recipes and their temporal evolution for improving customer satisfaction in tourism contexts. By extracting key service attributes and affects directly from tourist reviews, it identifies patterns that consistently achieve high satisfaction over time. This is the first study to integrate online review mining with panel fuzzy-set qualitative comparative analysis from a customer experience perspective. Findings reveal the existence of multiple equivalent causal pathways, with distinct satisfaction pathways for different tourist segments. Practically, normative causal recipes assist policymakers in optimizing tourism resource allocation and providing strategic insights to dynamically respond to tourist demands.

With the increasing digitization and networking of medical data and personal health information, information security has become a critical factor in vendor selection. However, limited understanding exists regarding how information security influences vendor selection. Drawing from the attention-based view (ABV), this study examines the potential impact of data breaches on hospitals' selection of electronic medical record system (EMRS) vendors. To test our hypotheses, we compile a unique dataset spanning 12 years of observations from US hospitals. Utilizing a coarsened exact matching (CEM) technique combined with a difference-in-differences (DiD) approach, our study shows that hospitals tend to replace their EMRS vendors after experiencing data breaches. Moreover, breached hospitals tend to prioritize information security in such a vendor replacement process by switching to star vendors and migrating towards a single-sourcing configuration. Further post-hoc analyses reveal that these impacts of data breaches are mitigated as the relationship between breached hospitals and vendors matures or when hospitals belong to large healthcare systems. Additionally, we find that the effects of data breaches are contingent on the scale of the breach and are short-term in nature. This research underscores the significance of information security as a crucial consideration in vendor selection for both academia and practitioners. We find that hospitals tend to change their electronic medical record system (EMRS) vendors following data breaches. During this replacement, they are more inclined to select star vendors and migrate towards a single-sourcing configuration. We also find that the impacts of data breaches can be mitigated as the relationship between breached hospitals and vendors matures, or when hospitals are integrated into larger healthcare systems. Additionally, we find that the effects of data breaches are dependent on the scale of the breach and are typically short-term in nature.

This study explores the effects of online behavior toward associated co-visited products on purchasing focal products. The effects are proposed to be moderated by brands, online reviews, and consumer experience. Empirical findings show that different dimensions of online behavior play distinctive roles in purchasing focal products. Furthermore, associated co-visited products across different brands have a significantly high effect. The effect also increases for associated co-visited products with low rates of negative reviews and for consumers with limited experience. This study contributes to the literature by highlighting the effects of online behavior and presenting the related implications for encouraging customer purchases.

This editorial introduces the special issue "The Interplay Between Artificial Intelligence, Production Systems, and Operations Management Resilience.' We selected twelve papers, encompassing many angles that illuminate the advances and challenges dealing with artificial intelligence tools and approaches in the production systems and operations management resilience domains. This editorial presents the papers with a smart view, highlighting the essentials of each article, such as full paper title, background, theory/literature scope, methodology design/analysis approach, and the main findings/contributions. Finally, the conclusions, future pathways, and research directions are presented.

We systematically review the data-driven innovation (DDI) literature spanning 2009-2022, analyzing key studies, assessing the current state of DDI research in information systems and operations management, and highlighting the research gaps. A classification framework to organize the DDI research literature is proposed. Using Gregor's (2006) theory classification framework, we identify theory types in the DDI literature. By applying the structural view (level of analysis) of Smith et al. (2011), we also investigate the level of analysis in the DDI literature. Our systematic literature review and analysis provide a roadmap both to facilitate knowledge creation and accumulation and to guide future DDI research. This review, the first of its kind focusing on DDI, summarizes DDI development, and identifies opportunities for new research, concluding with directions for future exploration in the field.

The rise of financial technology (fintech) has motivated practitioners and researchers to explore alternative data sources and enhanced credit scoring methods for better assessment of consumers’ credit risk. In this study, we examine whether deep-level diversity derived from consumers’ multimodal social media posts (i.e., alternative data) can enhance credit risk assessment or not. First, we propose novel lifestyle-based risk constructs (e.g., opinion risk) to capture consumers’ deep-level diversity. Second, we incorporate these lifestyle-based risk constructs into econometric models to empirically evaluate the relationship between consumers’ deep-level diversity and their credit risk. Using a credit scoring dataset provided by a fintech firm listed on Nasdaq, our econometric analysis reveals that consumers’ opinion risk constructs extracted from their multimodal social media posts are positively associated with their credit risk. Furthermore, our results show that the proposed opinion risk constructs can significantly improve the effectiveness of predicting consumers’ credit risk. Interestingly, our empirical results also show that combining the opinion risk constructs derived from images and text can significantly improve the effectiveness in credit risk prediction. This work contributes to the fintech domain by proposing novel lifestyle-based risk constructs for decision support in the credit scoring context.

With the rapid development of the capital market, financial fraud cases are becoming increasingly common. The evolving fraud strategies pose significant threats to financial regulation, market order, and the interests of ordinary investors. In order to combine the generalization performance of different machine learning methods and improve the effectiveness of financial fraud prediction, this paper proposes a novel financial fraud prediction framework based on stacking ensemble learning. This framework, based on data from listed companies, comprehensively considers financial ratio indicators and non-financial indicators. It uses the stacking ensemble technique to integrate numerous base models of machine learning algorithms for predicting financial fraud. Furthermore, the proposed framework has high versatility and is suitable for various tasks related to financial fraud prediction, addressing the problem of model selection difficulties in previous research due to different scenarios and data. We also conducted case studies on specific companies and industries, confirming the significant interpretability and practical applicability of the proposed framework. The results show that the recall rate and Area Under Curve (AUC) of our framework reached 0.8246 and 0.8146, respectively, surpassing mainstream machine learning models such as XGBoost and LightGBM in existing studies. This research study is of great significance for predicting the increasing number of financial fraud cases, providing a reliable tool for financial regulatory institutions and investors.

One effective solution for companies to expand their market share is to assign customers to optimal marketing strategy, also known as treatments, which are widely employed in randomized controlled trials or A/B tests. However, achieving optimal treatment assignment poses great challenges due to the limitations such as financial constraints and ethical issues associated with randomized controlled trials. To address the challenges, we propose a counterfactual-based uplift modeling approach. This approach involves generating counterfactual treatments and estimating corresponding effects using supervised learning models, ultimately determining the optimal treatment. Our methods have been evaluated on both synthetic and real-world data, demonstrating superior performance compared to other uplift modeling approaches in terms of the Qini coefficient. This study not only contributes to the research on causal inference in the business field but also offers practical implications for companies seeking to enhance business performance through effective marketing treatment assignment.

In recent years, there has been a growing emphasis on the role of innovation in enhancing resilience within supply chains. Furthermore, an increased interest in innovation as a key driver of resilience can be observed in practice. Despite this increased interest, research investigating the effects of supply chain innovation (SCI) on supply chain resilience (SCR) remains limited. To address this gap, we conceptualized a theoretical framework, grounded in the dynamic capabilities view, for testing the effect of SCI on SCR. Our research model further tested whether and how environmental uncertainty (EU) and top management involvement (TMI) moderate the effect of SCI on SCR, using structural equation modelling and survey data from 212 senior managers of firms in the textile and clothing industry in China. The study's findings offer a nuanced understanding of SCR and implications regarding SCI, EU, and TMI. Implications and suggestions for further research are also provided.