Pub Date : 2025-06-30DOI: 10.1016/j.dss.2025.114501
François de Corbière , Hirotoshi Takeda , Johanna Habib , Frantz Rowe , Daniel Thiel
To improve its economic and environmental performance, Carrefour, a major European retailer, restructured the distribution of logistic flows from its small and medium suppliers by introducing consolidation centers to expand flows and optimize resource sharing. The success of such an innovative supply chain (SC) largely depends on the number of suppliers deciding to adopt it without reverting to the previous SC. This specific context prompted us to propose a multi-agent model to analyze how the success of SC restructuring evolves as a function of delivery costs, information system (IS) integration and assimilation, and institutional pressures. Simulation results show first that, the lower IS integration in both the extant and the new SC, the more firms switch to and stay in the new SC. Second, a high level of IS assimilation in the new SC structure combined with coercive pressures fosters the success of SC restructuring.
{"title":"An agent-based model to analyze the influence of IS integration and IS assimilation on the adoption dynamics of a green supply chain: The case of regional consolidation centers","authors":"François de Corbière , Hirotoshi Takeda , Johanna Habib , Frantz Rowe , Daniel Thiel","doi":"10.1016/j.dss.2025.114501","DOIUrl":"10.1016/j.dss.2025.114501","url":null,"abstract":"<div><div>To improve its economic and environmental performance, Carrefour, a major European retailer, restructured the distribution of logistic flows from its small and medium suppliers by introducing consolidation centers to expand flows and optimize resource sharing. The success of such an innovative supply chain (SC) largely depends on the number of suppliers deciding to adopt it without reverting to the previous SC. This specific context prompted us to propose a multi-agent model to analyze how the success of SC restructuring evolves as a function of delivery costs, information system (IS) integration and assimilation, and institutional pressures. Simulation results show first that, the lower IS integration in both the extant and the new SC, the more firms switch to and stay in the new SC. Second, a high level of IS assimilation in the new SC structure combined with coercive pressures fosters the success of SC restructuring.</div></div>","PeriodicalId":55181,"journal":{"name":"Decision Support Systems","volume":"196 ","pages":"Article 114501"},"PeriodicalIF":6.7,"publicationDate":"2025-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144572371","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-06-28DOI: 10.1016/j.dss.2025.114498
Sen Yan, Zhiyi Wang, David Dobolyi
The recent development of generative AI (GenAI) algorithms has allowed machines to create new content in a realistic way, driving the spread of AI-generated content (AIGC) on the Internet. However, generative AI models and AIGC have exacerbated several societal challenges such as security threats (e.g., misinformation), trust issues, ethical concerns, and intellectual property regulation, calling for effective detection methods and a better understanding of AI-generated vs. human-written content. In this paper, we focus on AI-generated texts produced by large language models (LLMs) and extend prior detection methods by proposing a novel framework that combines semantic information and linguistic features. Based on potential semantic and linguistic differences in AI vs. human writing, we design our Semantic-Linguistic-Detector (SemLinDetector) framework by integrating a transformer-based semantic encoder and a linguistic encoder with parallel linguistic representations. By comparing a series of benchmark models on datasets collected from various LLMs and human writers in multiple domains, our experiments show that the proposed detection framework outperforms other benchmarks in a consistent and robust manner. Moreover, our model interpretability analysis showcases our framework's potential to help understand the reasoning behind prediction outcomes and identify patterns of differences in AI-generated and human-written content. Our research adds to the growing space of GenAI by proposing an effective and responsible detection system to address the risks and challenges of GenAI, offering implications for researchers and practitioners to better understand and regulate AIGC.
{"title":"An explainable framework for assisting the detection of AI-generated textual content","authors":"Sen Yan, Zhiyi Wang, David Dobolyi","doi":"10.1016/j.dss.2025.114498","DOIUrl":"10.1016/j.dss.2025.114498","url":null,"abstract":"<div><div>The recent development of generative AI (GenAI) algorithms has allowed machines to create new content in a realistic way, driving the spread of AI-generated content (AIGC) on the Internet. However, generative AI models and AIGC have exacerbated several societal challenges such as security threats (e.g., misinformation), trust issues, ethical concerns, and intellectual property regulation, calling for effective detection methods and a better understanding of AI-generated vs. human-written content. In this paper, we focus on AI-generated texts produced by large language models (LLMs) and extend prior detection methods by proposing a novel framework that combines semantic information and linguistic features. Based on potential semantic and linguistic differences in AI vs. human writing, we design our Semantic-Linguistic-Detector (SemLinDetector) framework by integrating a transformer-based semantic encoder and a linguistic encoder with parallel linguistic representations. By comparing a series of benchmark models on datasets collected from various LLMs and human writers in multiple domains, our experiments show that the proposed detection framework outperforms other benchmarks in a consistent and robust manner. Moreover, our model interpretability analysis showcases our framework's potential to help understand the reasoning behind prediction outcomes and identify patterns of differences in AI-generated and human-written content. Our research adds to the growing space of GenAI by proposing an effective and responsible detection system to address the risks and challenges of GenAI, offering implications for researchers and practitioners to better understand and regulate AIGC.</div></div>","PeriodicalId":55181,"journal":{"name":"Decision Support Systems","volume":"196 ","pages":"Article 114498"},"PeriodicalIF":6.7,"publicationDate":"2025-06-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144518575","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-06-27DOI: 10.1016/j.dss.2025.114500
Shweta Kumari Choudhary, Arpan Kumar Kar
In today's digital environment, organizations face security challenges like intentional breaches influenced by their specific policies and structures. As emerging technologies like Generative Artificial Intelligence (GAI) become more integrated into organizational processes, the adoption of GAI moderates organizational contextual conditions and rule characteristics, which affects the perceived risk of violating security rules. We extend the SOIPSV model to analyze cybersecurity practices and the strategic use of GAI in enhancing organizational resilience against security breaches. We establish the direct and moderating impacts of contextual conditions and rule characteristics, along with interactions in complex organizational cyber security. Our first study uses text mining for inferential and configurational analysis. Our second qualitative study explained the model of dynamic interplay between GAI and organizational factors. Our findings have implications for perceived risk management and managers redesigning business processes to manage security breaches.
{"title":"Modeling the role of generative AI in organizational privacy and security","authors":"Shweta Kumari Choudhary, Arpan Kumar Kar","doi":"10.1016/j.dss.2025.114500","DOIUrl":"10.1016/j.dss.2025.114500","url":null,"abstract":"<div><div>In today's digital environment, organizations face security challenges like intentional breaches influenced by their specific policies and structures. As emerging technologies like Generative Artificial Intelligence (GAI) become more integrated into organizational processes, the adoption of GAI moderates organizational contextual conditions and rule characteristics, which affects the perceived risk of violating security rules. We extend the SOIPSV model to analyze cybersecurity practices and the strategic use of GAI in enhancing organizational resilience against security breaches. We establish the direct and moderating impacts of contextual conditions and rule characteristics, along with interactions in complex organizational cyber security. Our first study uses text mining for inferential and configurational analysis. Our second qualitative study explained the model of dynamic interplay between GAI and organizational factors. Our findings have implications for perceived risk management and managers redesigning business processes to manage security breaches.</div></div>","PeriodicalId":55181,"journal":{"name":"Decision Support Systems","volume":"196 ","pages":"Article 114500"},"PeriodicalIF":6.7,"publicationDate":"2025-06-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144503653","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Category tree view is an omnipresent element in graphical user interfaces where it captures information in terms of a hierarchical structure. These categorization trees facilitate human users' cognitive economy and decision-making. While previous research has investigated the utilities of using unstructured data compared to pre-categorized information by business users, the effectiveness of allowing users the autonomy to create their own categorization hierarchies from generic object types remains unexplored. This paper evaluates the benefits of categorization autonomy in terms of search precision, as an objective measure, as well as subjective intentions to use the system. We examined users' interactions with a platform in information seeking tasks with 201 subjects. Our findings indicate that categorization autonomy leads to superior results, both in terms of effective use and behavioral perceptions. We also found that the impact of categorization autonomy is moderated by task flexibility, such that the benefits are more apparent in tasks that necessitate open-ended search approaches. By focusing on how user-driven categorization influences system interaction, our study contributes to the design of decision support systems that are better aligned with users' cognitive structures and task demands.
{"title":"Impact of categorization autonomy on effective use and adoption intentions","authors":"Arash Saghafi , Poonacha Medappa , Ariton Debrliev","doi":"10.1016/j.dss.2025.114499","DOIUrl":"10.1016/j.dss.2025.114499","url":null,"abstract":"<div><div>Category tree view is an omnipresent element in graphical user interfaces where it captures information in terms of a hierarchical structure. These categorization trees facilitate human users' cognitive economy and decision-making. While previous research has investigated the utilities of using unstructured data compared to pre-categorized information by business users, the effectiveness of allowing users the autonomy to create their own categorization hierarchies from generic object types remains unexplored. This paper evaluates the benefits of categorization autonomy in terms of search precision, as an objective measure, as well as subjective intentions to use the system. We examined users' interactions with a platform in information seeking tasks with 201 subjects. Our findings indicate that categorization autonomy leads to superior results, both in terms of effective use and behavioral perceptions. We also found that the impact of categorization autonomy is moderated by task flexibility, such that the benefits are more apparent in tasks that necessitate open-ended search approaches. By focusing on how user-driven categorization influences system interaction, our study contributes to the design of decision support systems that are better aligned with users' cognitive structures and task demands.</div></div>","PeriodicalId":55181,"journal":{"name":"Decision Support Systems","volume":"196 ","pages":"Article 114499"},"PeriodicalIF":6.7,"publicationDate":"2025-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144503391","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-06-24DOI: 10.1016/j.dss.2025.114497
Arslan Rafi , Sanjit K. Roy , Mohsin Abdur Rehman , Muhammad Junaid Shahid Hasni
In the original article, we examined various factors, including social, spatial, and self-presence, influencing user well-being in metaverse communities. We intended to examine the symmetrical and asymmetrical relationships between types of presence and user well-being. However, discrepancies emerged in reporting the final measurement items and their validity assessment. We provide details on how we corrected the errors in the article.
{"title":"Corrigendum to “Impact of multidimensional presence on user well-being in metaverse communities”","authors":"Arslan Rafi , Sanjit K. Roy , Mohsin Abdur Rehman , Muhammad Junaid Shahid Hasni","doi":"10.1016/j.dss.2025.114497","DOIUrl":"10.1016/j.dss.2025.114497","url":null,"abstract":"<div><div>In the original article, we examined various factors, including social, spatial, and self-presence, influencing user well-being in metaverse communities. We intended to examine the symmetrical and asymmetrical relationships between types of presence and user well-being. However, discrepancies emerged in reporting the final measurement items and their validity assessment. We provide details on how we corrected the errors in the article.</div></div>","PeriodicalId":55181,"journal":{"name":"Decision Support Systems","volume":"196 ","pages":"Article 114497"},"PeriodicalIF":6.7,"publicationDate":"2025-06-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144503547","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-06-22DOI: 10.1016/j.dss.2025.114496
Yu Han , Ziqiong Zhang , Carol X.J. Ou , Zili Zhang
Review helpfulness is crucial for assessing the quality of online reviews and mitigating information overload. Although numerous studies have explored the impact of textual and reviewer characteristics on review helpfulness, the role of photo aesthetics remains important but underexplored. This study addresses this gap by investigating the impact of photo aesthetics on perceived review helpfulness and its underlying mediating effects. The hotel review data from TripAdvisor.com exhibit an inverted U-shaped effect of photo aesthetics on perceived review helpfulness, in which review text length moderates this relationship. To further validate this causal relationship and explore the underlying mediating effects, an experimental study is conducted. The experimental results confirm the causal impact of photo aesthetics on perceived review helpfulness and reveal that perceived pleasure, reviewer effort and review authenticity mediate the relationship. These novel insights challenge the notion that “the more aesthetic, the better” for review photos, offering new theoretical and practical implications.
{"title":"The more aesthetic, the better? The impact of photo aesthetics on perceived review helpfulness","authors":"Yu Han , Ziqiong Zhang , Carol X.J. Ou , Zili Zhang","doi":"10.1016/j.dss.2025.114496","DOIUrl":"10.1016/j.dss.2025.114496","url":null,"abstract":"<div><div>Review helpfulness is crucial for assessing the quality of online reviews and mitigating information overload. Although numerous studies have explored the impact of textual and reviewer characteristics on review helpfulness, the role of photo aesthetics remains important but underexplored. This study addresses this gap by investigating the impact of photo aesthetics on perceived review helpfulness and its underlying mediating effects. The hotel review data from <span><span>TripAdvisor.com</span><svg><path></path></svg></span> exhibit an inverted U-shaped effect of photo aesthetics on perceived review helpfulness, in which review text length moderates this relationship. To further validate this causal relationship and explore the underlying mediating effects, an experimental study is conducted. The experimental results confirm the causal impact of photo aesthetics on perceived review helpfulness and reveal that perceived pleasure, reviewer effort and review authenticity mediate the relationship. These novel insights challenge the notion that “the more aesthetic, the better” for review photos, offering new theoretical and practical implications.</div></div>","PeriodicalId":55181,"journal":{"name":"Decision Support Systems","volume":"196 ","pages":"Article 114496"},"PeriodicalIF":6.7,"publicationDate":"2025-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144502471","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-06-20DOI: 10.1016/j.dss.2025.114495
Abhipsa Pal , Rahul Dé , H. Raghav Rao
Although the diffusion of mobile payment technology has been historically governed by contextual events that trigger anxiety, accentuating either the risks of mobile payments or the risks of its conflicting alternative, cash, literature neglects the importance of examining the risks associated with the alternatives. To address this gap, we develop the risk-risk trade-off (R2T) framework, drawing from the theory of substitutes of hazardous substances, and examine how individuals make usage decisions by balancing two sets of risks – for mobile payments and cash, respectively. On one side, the framework weighs contactless [mobile payment] risks related to potential thefts and losses, heightened by the rise in cybercrime. Conversely, on the other side, it weighs the risks from its substitute, contact [cash] payment, carrying the health hazard of infectious disease transmission through contact, with this risk magnified during the global pandemic. To validate the model, we used survey responses from 1403 participants in India and triangulated the quantitative results using their qualitative comments. This study theoretically contributes to the mobile payment usage literature by moving beyond technology risks as the sole risks to be considered for usage decision-making and includes the analysis of risks of the technology's substitute, cash, as well. The framework can support analysis of users' decisions towards consciously choosing the technology against its alternatives, in various risky contexts.
{"title":"The risk-risk trade-off (R2T) framework: Examining contact [cash] versus contactless [mobile] payment usage","authors":"Abhipsa Pal , Rahul Dé , H. Raghav Rao","doi":"10.1016/j.dss.2025.114495","DOIUrl":"10.1016/j.dss.2025.114495","url":null,"abstract":"<div><div>Although the diffusion of mobile payment technology has been historically governed by contextual events that trigger anxiety, accentuating either the risks of mobile payments or the risks of its conflicting alternative, cash, literature neglects the importance of examining the risks associated with the alternatives. To address this gap, we develop the risk-risk trade-off (R<sup>2</sup>T) framework, drawing from the theory of substitutes of hazardous substances, and examine how individuals make usage decisions by balancing two sets of risks – for mobile payments and cash, respectively. On one side, the framework weighs contactless [mobile payment] risks related to potential thefts and losses, heightened by the rise in cybercrime. Conversely, on the other side, it weighs the risks from its substitute, contact [cash] payment, carrying the health hazard of infectious disease transmission through contact, with this risk magnified during the global pandemic. To validate the model, we used survey responses from 1403 participants in India and triangulated the quantitative results using their qualitative comments. This study theoretically contributes to the mobile payment usage literature by moving beyond technology risks as the sole risks to be considered for usage decision-making and includes the analysis of risks of the technology's substitute, cash, as well. The framework can support analysis of users' decisions towards consciously choosing the technology against its alternatives, in various risky contexts.</div></div>","PeriodicalId":55181,"journal":{"name":"Decision Support Systems","volume":"196 ","pages":"Article 114495"},"PeriodicalIF":6.7,"publicationDate":"2025-06-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144335630","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The rise of Location-based Social Networking (LBSN) platforms has transformed the way users explore Points of Interest (POIs), increasingly relying on group-based recommendations. However, recommending POIs to groups presents unique challenges due to conflicting preferences among members. Traditional group recommendation algorithms often prioritize aggregated methods or explicit preference extraction, overlooking latent domain-specific information and the dynamic nature of group decision-making. To address these gaps, we propose a novel Knowledge-based Context-Aware Group Recommender System (KCGRS) designed to support decision-making processes within groups. KCGRS operates in two key stages: first, it utilizes a knowledge graph to learn domain-specific embeddings for both users and POIs, ensuring that implicit preferences and contextual factors are incorporated. In the second stage, these embeddings are enhanced with contextual information using a feed-forward transformer model, allowing for a more nuanced understanding of real-time preferences. The decision-making process is further refined by generating a group embedding, which is computed by applying a weighted aggregate of the context-infused embeddings of individual group members. This approach models group dynamics and decision processes more accurately, ensuring that the final recommendation reflects the collective preferences of the group. Experiments on real-world Yelp data show that KCGRS significantly outperforms five state-of-the-art baselines, delivering up to an average of 14.15% improvement in Hit ratio and a 13.07% increase in NDCG compared to the next best method while also maintaining competitive runtime efficiency. Furthermore, KCGRS demonstrates enhanced diversity and coverage in recommendations, ensuring that POI suggestions cater to a broader range of user preferences while avoiding over-personalization. This balance between accuracy, diversity, and efficiency highlights KCGRS’s effectiveness in supporting group decision-making and its potential to enhance collaborative recommendations in LBSN platforms. Finally, a user study with 144 participants was conducted that resulted in statistically significant levels of user satisfaction and trust in the recommendations, thereby supporting the practical effectiveness of the KCGRS system.
{"title":"Knowledge-based Context-aware Group Recommender System for Point of Interest recommendation","authors":"Nargis Pervin , Abhishek Kulkarni , Ayush Adarsh , Shreya Som","doi":"10.1016/j.dss.2025.114485","DOIUrl":"10.1016/j.dss.2025.114485","url":null,"abstract":"<div><div>The rise of Location-based Social Networking (LBSN) platforms has transformed the way users explore Points of Interest (POIs), increasingly relying on group-based recommendations. However, recommending POIs to groups presents unique challenges due to conflicting preferences among members. Traditional group recommendation algorithms often prioritize aggregated methods or explicit preference extraction, overlooking latent domain-specific information and the dynamic nature of group decision-making. To address these gaps, we propose a novel Knowledge-based Context-Aware Group Recommender System (KCGRS) designed to support decision-making processes within groups. KCGRS operates in two key stages: first, it utilizes a knowledge graph to learn domain-specific embeddings for both users and POIs, ensuring that implicit preferences and contextual factors are incorporated. In the second stage, these embeddings are enhanced with contextual information using a feed-forward transformer model, allowing for a more nuanced understanding of real-time preferences. The decision-making process is further refined by generating a group embedding, which is computed by applying a weighted aggregate of the context-infused embeddings of individual group members. This approach models group dynamics and decision processes more accurately, ensuring that the final recommendation reflects the collective preferences of the group. Experiments on real-world Yelp data show that KCGRS significantly outperforms five state-of-the-art baselines, delivering up to an average of 14.15% improvement in Hit ratio and a 13.07% increase in NDCG compared to the next best method while also maintaining competitive runtime efficiency. Furthermore, KCGRS demonstrates enhanced diversity and coverage in recommendations, ensuring that POI suggestions cater to a broader range of user preferences while avoiding over-personalization. This balance between accuracy, diversity, and efficiency highlights KCGRS’s effectiveness in supporting group decision-making and its potential to enhance collaborative recommendations in LBSN platforms. Finally, a user study with 144 participants was conducted that resulted in statistically significant levels of user satisfaction and trust in the recommendations, thereby supporting the practical effectiveness of the KCGRS system.</div></div>","PeriodicalId":55181,"journal":{"name":"Decision Support Systems","volume":"196 ","pages":"Article 114485"},"PeriodicalIF":6.7,"publicationDate":"2025-06-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144322516","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-06-16DOI: 10.1016/j.dss.2025.114493
Michael Hagn, Bernd Heinrich, Thomas Krapf, Alexander Schiller
In recent years, machine learning (ML) has become ubiquitous in sectors including transportation, security, health, and finance to analyze large amounts of data and support decision-making. However, real-world datasets used in ML often exhibit various data quality (DQ) defects that can significantly impair the performance and validity of ML models and thus also the decisions derived from them. Therefore, a plethora of methods across various research strands have been proposed to address DQ defects and mitigate their negative impact on ML-based data analysis and decision support. This has resulted in a fragmented research landscape, where comparisons and classifications of methods dealing with ML on data with DQ defects are very challenging for both researchers and practitioners. Thus, based on a structured design process, we develop and present a taxonomy for this research field. The taxonomy serves as a systematic framework to classify and organize existing research and methods according to relevant dimensions and facilitates future work in this area. Its reliability, understandability, completeness, and usefulness are supported by an evaluation with external researchers and practitioners. Finally, we identify current trends and research gaps and derive challenges and directions for future research.
{"title":"Handling imperfection: A taxonomy for machine learning on data with data quality defects","authors":"Michael Hagn, Bernd Heinrich, Thomas Krapf, Alexander Schiller","doi":"10.1016/j.dss.2025.114493","DOIUrl":"10.1016/j.dss.2025.114493","url":null,"abstract":"<div><div>In recent years, machine learning (ML) has become ubiquitous in sectors including transportation, security, health, and finance to analyze large amounts of data and support decision-making. However, real-world datasets used in ML often exhibit various data quality (DQ) defects that can significantly impair the performance and validity of ML models and thus also the decisions derived from them. Therefore, a plethora of methods across various research strands have been proposed to address DQ defects and mitigate their negative impact on ML-based data analysis and decision support. This has resulted in a fragmented research landscape, where comparisons and classifications of methods dealing with ML on data with DQ defects are very challenging for both researchers and practitioners. Thus, based on a structured design process, we develop and present a taxonomy for this research field. The taxonomy serves as a systematic framework to classify and organize existing research and methods according to relevant dimensions and facilitates future work in this area. Its reliability, understandability, completeness, and usefulness are supported by an evaluation with external researchers and practitioners. Finally, we identify current trends and research gaps and derive challenges and directions for future research.</div></div>","PeriodicalId":55181,"journal":{"name":"Decision Support Systems","volume":"196 ","pages":"Article 114493"},"PeriodicalIF":6.7,"publicationDate":"2025-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144338972","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-06-16DOI: 10.1016/j.dss.2025.114492
Xinwei Wang , Yi Feng , Sutong Wang , Dujuan Wang , T.C.E. Cheng
Utilizing machine learning methods for auxiliary decision support in medical imaging significantly reduces missed detections and unnecessary expenses. However, the strict accuracy and transparency requirements in the medical field pose challenges for deep learning applications based on neural networks. To address these issues, we propose a novel artificial intelligence artifact guided by the design science research methodology for lesion detection decision support in medical images, called Explainable Lesion DEtection TRansformer (EL-DETR). This approach features an explainable separate attention mechanism in the decoder that highlights the attention weights of content and location queries, providing insights into the inference process through attention mapping visualizations. In addition, we introduce a hybrid matching query strategy to enhance the learning of positive samples and develop an adaptive efficient compound loss function to optimize training. We demonstrate EL-DETR's superior accuracy, robustness, and interpretability using four real-world datasets, establishing it as a reliable tool for clinical diagnosis and treatment decision support based on medical imaging. The code and original data are available at https://github.com/weimingai/EL-DETR.
{"title":"An explainable lesion detection transformer model for medical imaging diagnosis decision support: Design science research","authors":"Xinwei Wang , Yi Feng , Sutong Wang , Dujuan Wang , T.C.E. Cheng","doi":"10.1016/j.dss.2025.114492","DOIUrl":"10.1016/j.dss.2025.114492","url":null,"abstract":"<div><div>Utilizing machine learning methods for auxiliary decision support in medical imaging significantly reduces missed detections and unnecessary expenses. However, the strict accuracy and transparency requirements in the medical field pose challenges for deep learning applications based on neural networks. To address these issues, we propose a novel artificial intelligence artifact guided by the design science research methodology for lesion detection decision support in medical images, called Explainable Lesion DEtection TRansformer (EL-DETR). This approach features an explainable separate attention mechanism in the decoder that highlights the attention weights of content and location queries, providing insights into the inference process through attention mapping visualizations. In addition, we introduce a hybrid matching query strategy to enhance the learning of positive samples and develop an adaptive efficient compound loss function to optimize training. We demonstrate EL-DETR's superior accuracy, robustness, and interpretability using four real-world datasets, establishing it as a reliable tool for clinical diagnosis and treatment decision support based on medical imaging. The code and original data are available at <span><span>https://github.com/weimingai/EL-DETR</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":55181,"journal":{"name":"Decision Support Systems","volume":"196 ","pages":"Article 114492"},"PeriodicalIF":6.7,"publicationDate":"2025-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144291619","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号