Decision Support Systems最新文献

Online retail platforms encounter numerous challenges, such as cyber-attacks, data breaches, device failures, and operational disruptions. These challenges have intensified in recent years, underscoring the importance of prioritizing resilience for businesses. Unfortunately, conventional cybersecurity methods have proven insufficient in thwarting sophisticated cybercrime tactics. This paper proposes a novel resilience strategy that leverages Explainable Deep Learning technologies and a Blockchain-based consensus protocol strategy. By combining these two approaches, our strategy enables rapid incident detection, explains the features and related vulnerabilities that are used, and enhances decision-making during cyber incidents. To validate the efficacy of our approach, we conducted experiments using NAB datasets, preprocessed and trained the data, and performed an experimental study on real online retail architectures. Our results demonstrate the effectiveness of the proposed framework in supporting business and operation continuity and creating more efficient cyber resilience strategies that will enhance decision-making capabilities.

Improving customer repurchase intention constitutes a key activity for maintaining sustainable business performance. Returning customers provide many economic and other benefits to businesses. In contrast, attracting new customers is a process that is associated with high costs. This work proposes a novel counterfactual explanations methodology that utilizes textual data from electronic word of mouth to recommend business changes that can improve customers' repurchase behavior. Counterfactual explanation methods gained considerable attention because their logic aligns with human reasoning and the fact that they can recommend low-cost actions on how to turn an unfavorable outcome into a favorable. Most counterfactual explanation methods however recommend actions that can change the outcome of individual instances (i.e. one customer) rather than a group of instances. Therefore, this work proposes a multi-instance counterfactual explanation method that recommends optimum changes to an organization's practices/policies that increase repurchase intention for many customers or customer segments.

The proposed methodology utilizes topic modeling to extract customer opinions from online reviews' text and use topics as features to train a binary classifier that predicts customer revisit intention. Multi-instance counterfactual explanations are computed for all or different groups of non-revisiting customers, recommending optimum business changes that can increase revisit intention. The proposed methodology is empirically evaluated through a case study on the restaurant revisit problem and compared against a prominent alternative from the literature. The results show that the method has better performance to the alternative method and produces recommendations that are actionable and abide by the customer-repurchase literature.

In an era of growing social activism, companies engaged in socially irresponsible practices are increasingly vulnerable to data breaches, resulting in substantial reputational and financial losses. This study examines how corporate social irresponsibility (CSI) influences a company's data breach risk. We argue that CSI has an impact on data breach risk by influencing the intentional behaviors of both employees and external hackers. Given that CSI is a broad concept and can take on various forms, we further examine whether some forms of CSI pose a more significant threat than others. Our empirical analysis of data breaches in publicly listed US firms from 2005 to 2017 indicates that compared to the forms of CSI that violate broader social norms (e.g., environmental damages), CSI activities that jeopardize a company's economic value delivery (e.g., product deficiencies) play a more dominant role in driving data breach risk. Furthermore, we find that corporate social responsibility (CSR) can have a dual impact on moderating the relationship between CSI and data breaches. While CSR often helps mitigate CSI-induced data breach risk, this risk is heightened when both CSR and CSI relate to a firm's economic value delivery. This study provides critical insights into how companies can navigate complex data breach risk by managing their social performance.

Since the Economist magazine heralded blockchain as “the trust machine” in 2015, the blockchain paradigm has experienced crests and falls, including a recent phase of disillusionment due to its failure to meet the high expectations, e.g., to revolutionize record keeping, data management, and workflow, envisioned during its early history. However, despite the waning interest in this technology in some quarters, its deployment has become ever more essential in areas such as decentralized finance (DeFi), Non-fungible Tokens (NFTs), and other application domains beyond cryptocurrencies. In particular, recent advancements in Artificial Intelligence (AI) surrounding Large Language Models (LLM) offer new opportunities for blockchain adoption where trust and reliability become critical. As the blockchain technology transitions from a stage of disillusionment to one of enlightenment, anticipation is building for its mainstream adoption, with focused endeavors towards removing adoption barriers across diverse business contexts, exemplified by studies included in this special issue on Blockchain Technology and Applications. In this paper, we first survey the current state of the blockchain technology and then highlight its potential for enhancing trust and accountability in emerging phenomena such as AI generated content (AIGC). We conclude by introducing the papers included in the special issue.

The significance of online communities in our lives is indisputable. These communities take various forms, including social networking sites, brand communities, and virtual platforms, where individuals digitally connect and interact. This article suggests that users' perceptions and beliefs about online communities are shaped by multiple selection mechanisms, which significantly influence decision-making processes related to community participation. This article is supported by two studies, with the second study building upon the first. Study 1 retrospectively explores selection mechanisms by drawing from network theory, social capital theory, and motivation theory. Through principal component analysis, these mechanisms are identified and categorized as community selection mechanisms. In Study 2, the focus shifts to examining whether these mechanisms lead to differences in community engagement behaviors. These behaviors encompass intentions to continue participating, knowledge sharing, and electronic word-of-mouth (e-WOM). By comparing various communities based on their characteristics, the results reveal that each selection mechanism holds varying degrees of importance in influencing community engagement. For instance, content gratification is a key mechanism for the selection of professional and travel communities, but it lacks significance as a predictor for the game community. These findings not only advance our understanding of community selection mechanisms but also provides valuable insights for businesses looking to optimize their decision-making processes.

In our work, we propose the use of Representational Similarity Analysis (RSA) for explainable AI (XAI) approaches to enhance the reliability of XAI-based decision support systems. To demonstrate how similarity analysis of explanations can assess the output stability of post-hoc explainers, we conducted a computational evaluative study. This study addresses how our approach can be leveraged to analyze the stability of explanations amidst various changes in the ML pipeline. Our results show that modifications such as altered preprocessing or different ML models lead to changes in the explanations and illustrate the extent to which stability can suffer. Explanation similarity analysis enables practitioners to compare different explanation outcomes, thus monitoring stability in explanations. Alongside discussing the results and practical applications in operationalized ML, including both benefits and limitations, we also delve into insights from computational neuroscience and neural information processing.

Brick-and-mortar shopping malls are embracing Artificial Intelligence (AI) technology and recommender systems to enhance the shopping experience and boost mall revenue. Echoing this trend, we formulate a new shopping trip recommendation problem, which aims to recommend a shopping trip (i.e., a list of stores) that matches customer preferences and has appropriate trip lengths. To solve this problem, we develop a novel deep learning-enhanced global planning (DeepGP) approach featuring three methodological novelties. First, we introduce a new shopping intensity term based on deep neural networks to capture the variation of trip lengths specific to different shopping contexts. Second, we innovatively formulate the learning and optimization objectives in a consistent form by balancing the shopping choice likelihood and the shopping intensity likelihood, thus resolving the inconsistency issue encountered by prior global planning methods. Third, to overcome the computational challenge caused by the nonlinear shopping intensity term, we design a new exact and efficient solution technique based on piecewise linear transformations. Using a real-world offline shopping dataset, we empirically demonstrate the superior performances of our approach compared to representative benchmarks in offering more accurate and relevant shopping trip recommendations. Through a simulation, we show the capacity of our approach to attract and balance customer traffic in practical deployments. Overall, our research highlights the efficacy of combining shopping choices and shopping intensity in a consistent learning and optimization framework for offline shopping trip recommendations.

Creators have long strived to secure royalties for their works but with little success. In the digital realm, monetization presents an even greater challenge, as traditional digital assets frequently suffer from piracy issues, primarily due to the lack of verifiable ownership. Recently, non-fungible token (NFT), a blockchain-enabled tradable digital asset, has aroused great public attention for its potential to address this long-standing issue. Specifically, NFTs empower creators by enabling them to earn resale royalties from post-primary-sale transactions and by providing verifiable ownership that facilitates consumer trading in a secondary market. In this paper, we employ a two-stage game-theoretic model to examine this innovative business model. Here, an NFT creator makes optimal pricing and royalty rate decisions, while consumers make their purchase and resale decisions accordingly. Our study reveals that the creator inevitably reduces the selling price when a secondary market is introduced, even without imposing royalties. Moreover, contrary to conventional wisdom that NFT-enabled royalties always benefit creators, our research uncovers that the introduction of a secondary market leads to unintended revenue loss for the creator, particularly when most consumers only engage in secondary transactions. Furthermore, we find that the platform's profitability diminishes with the introduction of a secondary market, especially when most consumers are uninformed and the platform relies primarily on the primary market for commission collection. Finally, we find that the introduction of a secondary market may leave consumers worse off, despite the resale opportunities it offers. Our findings carry crucial managerial implications for platforms, creators, consumers, and policymakers.

We explore the question of skill versus chance dominance in Daily Fantasy Sports (DFS), which has been the subject of numerous legal disputes around the world. Our study examines whether a contestant's winnability in DFS is influenced by factors reflecting skills using cricket-based daily fantasy contest data and a true fixed effects stochastic frontier model. We find that skill contributes significantly towards winnability in five ways. First, contestants performing well in the past do better in the present. Second, gaining more game experiences improves performance. Third, contestants who participated recently, tend to exhibit higher winnability. Fourth, selecting an appropriate contest type enhances winnability. Fifth, the large estimated signal-to-noise ratio indicates that the unobserved skill measured by a non-negative error has a much greater impact on winnability than the regular two-sided random shocks. These results are robust to varying specifications and subsets of data. Decision makers and regulators can use the model presented in the study to distinguish skill-dominant DFS from chance-dominant DFS.

This study examines the impact of doctors' facial attractiveness on users' choices in online health communities (OHCs). We conducted a field study using a sample of 14,897 doctors registered on a Chinese OHC. The results indicate a significant negative relationship between the facial attractiveness of doctors and the number of visits to their homepage by users. However, this relationship only holds true for male surgeons and female internal medicine doctors, not for female surgeons or male internal medicine doctors. These findings suggest the possible presence of gender-specialty bias in the influence of facial attractiveness on patients' decision-making. To further investigate how facial attractiveness influences users' inclination to choose a particular doctor, we develop our research model drawing upon the stereotype content and social role perspective. Through a laboratory experiment, we found that OHC users' perceptions of doctors' warmth and competence act as mediating factors in the relationship between facial attractiveness and user choice. Additionally, this relationship is influenced by stereotypical gender-specialty fit.