WIREs Data Mining and Knowledge Discovery最新文献

{"title":"Challenges and Transformations in Biomedical Statistics Research Driven by Artificial Intelligence","authors":"Yang Liu, Ying Liu, Wen‐rong Zhao, Guo‐liang Fan, Hai‐tao Zhang","doi":"10.1002/widm.70121","DOIUrl":"https://doi.org/10.1002/widm.70121","url":null,"abstract":"With the rapid advancement of artificial intelligence (AI) technology, the field of biomedical statistics has embraced novel solutions in areas such as data processing, variable control, and model formulation. This paper explores how AI technology is driving a transformation in the research design framework and conceptual paradigms of biomedical statistics. The focus is, particularly, on the challenges faced by AI‐driven causal inference research, including issues with the misinterpretation of associations, data noise interference, and shortages of methods. To address these challenges, this paper proposes a solution strategy centered on transparent algorithm design, the integration of heterogeneous models, and third‐party validation mechanisms. These strategies aim to provide a methodological foundation for bridging the gap between AI‐driven statistical associations and biological mechanisms. This article is categorized under: <jats:list list-type=\"simple\"> <jats:list-item> Algorithmic Development &gt; Statistics </jats:list-item> <jats:list-item> Technologies &gt; Artificial Intelligence </jats:list-item> </jats:list>","PeriodicalId":501013,"journal":{"name":"WIREs Data Mining and Knowledge Discovery","volume":"35 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-08-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148675248","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Benchmarking Robustness in Copy‐Move Forgery Detection: A Comparative Review of Algorithmic Trade‐Offs and Transformation Resilience","authors":"Seema Narvare, Kunj Bihari Meena","doi":"10.1002/widm.70118","DOIUrl":"https://doi.org/10.1002/widm.70118","url":null,"abstract":"Copy‐move forgery detection (CMFD) remains a critical challenge in image forensics due to the increasing sophistication of post‐processing and geometric transformations. This paper presents a comprehensive benchmarking study and systematic review of CMFD research spanning the last decade (2015–2025). The study provides a qualitative synthesis of 91 empirical studies, categorized into four primary taxonomies: block‐based (36), keypoint‐based (27), deep learning‐based (17), and hybrid (11) methods. The study evaluates these approaches against realistic distortions, including JPEG compression, noise addition, blurring, and geometric operations such as rotation and scaling. The comparative analysis highlights significant trade‐offs among methodological categories: deep learning approaches demonstrate strong feature‐learning capabilities and robustness across diverse forgery scenarios, keypoint‐based methods offer a practical balance between computational efficiency and detection accuracy, while block‐based methods often achieve high localization precision at the expense of increased computational complexity. Furthermore, the performance of state‐of‐the‐art techniques is analyzed using F1‐measure results reported across widely used benchmark datasets. The review also discusses the challenges associated with cross‐study comparisons arising from variations in datasets, evaluation protocols, and reporting practices. Finally, this work identifies critical gaps in the field, including inconsistent benchmarking standards and dataset limitations, and outlines emerging research directions to support the development of more robust and generalizable image authentication systems.","PeriodicalId":501013,"journal":{"name":"WIREs Data Mining and Knowledge Discovery","volume":"30 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148633834","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Role of Human‐Centric Approach Toward Digital Sovereignty: A Systematic Literature Review","authors":"Tarun Kumar Behera, Sanjay Misra, Biswajeeban Mishra, Petter Kvalvik","doi":"10.1002/widm.70105","DOIUrl":"https://doi.org/10.1002/widm.70105","url":null,"abstract":"With the growing mediation of personal data and identity by digital technologies—driven not only by the significant growth of these technologies, but also by the widespread processing of personal data—digital sovereignty has emerged as a significant issue. Conventional methodologies prioritize national governance, geopolitical considerations, and centralized management of data infrastructures. Conversely, human‐centric digital sovereignty emphasizes individual rights, data agency, privacy protection, and community autonomy. Frameworks like the General Data Protection Regulation (GDPR) illustrate this transition by providing individuals substantial authority over personal data and enhancing digital trust and security. This study analyzes the integration of human‐centric concepts into legislative and technological frameworks to promote digital sovereignty. The focus encompasses: (1) distinctions between human‐centric and conventional sovereignty models; (2) the empowerment of individuals and communities via data ownership, privacy, and autonomy; (3) mechanisms facilitating implementation; (4) obstacles across technical, regulatory, and societal dimensions; and (5) opportunities and risks associated with emerging technologies such as artificial intelligence, blockchain, and the Internet of Things. A systematic review was performed using PRISMA principles across seven prominent academic databases: ACM Digital Library, IEEE Xplore, ScienceDirect, Wiley Online Library, Springer Link, Taylor &amp; Francis, and Google Scholar. Out of 5684 initial articles (2020–2025), a multi‐stage screening process discovered 70 peer‐reviewed research articles in English pertaining to digital sovereignty and human‐centric methodologies. Findings demonstrate a global transition from state‐centric to individual‐centric sovereignty. Decentralized technologies, especially Distributed Ledger Technologies (DLT), present significant opportunities for personal data governance, secure identity management, and international compliance; nonetheless, challenges such as scalability constraints, regulatory fragmentation, interoperability issues, and opposition from centralized organizations persist as substantial obstacles. Artificial intelligence, DLT technologies, and the Internet of Things offer privacy‐enhancing computation and decentralized identities; yet they also introduce concerns such as algorithmic bias, security vulnerabilities, and governance ambiguities. The future of digital sovereignty hinges on the prioritization of human agency, privacy, and community empowerment. Attaining holistic human‐centric sovereignty necessitates synchronized global regulation, ethical and security benchmarks, interoperable frameworks, and collaborative engagement across governments, industry, academia, and civil society. This article is categorized under: &lt;jats:list list-type=\"simple\"&gt; &lt;jats:list-item&gt; Application Areas &gt; Science and Technology &lt;/jats:list-item&gt; &lt;jats:list-item&gt; Commerci","PeriodicalId":501013,"journal":{"name":"WIREs Data Mining and Knowledge Discovery","volume":"43 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148633936","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Systematic Review and Taxonomy of Fall Detection in Geriatric Care: Current State, Challenges, and Applications","authors":"Hamid Ali, Ji Zhang, Xiaohui Tao, Raj Gururajan, Xujuan Zhou, Yuquan Gan, Josh Jia‐Ching Ying","doi":"10.1002/widm.70115","DOIUrl":"https://doi.org/10.1002/widm.70115","url":null,"abstract":"This systematic review provides a critical examination of fall detection (FD) technologies in geriatric care, analyzing both technological innovations and implementation challenges across diverse healthcare environments. The study employs a comprehensive evaluation framework to assess various technological approaches, which consists of wearable sensors and Internet of Things (IoT)‐enabled devices to machine learning (ML) algorithms and video monitoring systems, thereby identifying significant gaps in current research and practice. Despite the demonstrated improvements in detection accuracy and real‐time monitoring capabilities of contemporary systems, persistent challenges emerge in areas of system longevity, user acceptance, and healthcare system integration. Particularly concerning are the ethical implications surrounding privacy preservation and algorithmic transparency in AI‐driven systems, combined with inadequate consideration of demographic variations in fall patterns across different user populations. The analysis reveals an urgent need for developing more personalized, interpretable, and user‐centric FD solutions that address these multifaceted challenges. This work provides evidence‐based recommendations for future research, with particular emphasis on developing robust, ethically sound systems that integrate seamlessly with existing healthcare workflows. The research contributions advance the field through actionable insights for developing more effective and inclusive FD technologies that enhance the safety and independence of elderly individuals without compromising their privacy and autonomy. This article is categorized under: <jats:list list-type=\"simple\"> <jats:list-item> Application Areas &gt; Health Care </jats:list-item> <jats:list-item> Application Areas &gt; Science and Technology </jats:list-item> <jats:list-item> Technologies &gt; Machine Learning </jats:list-item> </jats:list>","PeriodicalId":501013,"journal":{"name":"WIREs Data Mining and Knowledge Discovery","volume":"271 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-07-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148633937","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Managing Diabetic Retinopathy With Deep Learning: A Data Centric Overview","authors":"Shramana Dey, Zahir Khan, T. A. PramodKumar, B. Uma Shankar, Ashis K. Dhara, Ramachandran Rajalakshmi, Rajiv Raman, Sushmita Mitra","doi":"10.1002/widm.70113","DOIUrl":"https://doi.org/10.1002/widm.70113","url":null,"abstract":"Diabetic retinopathy (DR) is a serious microvascular complication of diabetes, and one of the leading causes of vision loss worldwide. Although automated detection and grading, with deep learning (DL), can reduce the burden on ophthalmologists, it is constrained by the limited availability of high‐quality datasets. Existing repositories often remain geographically narrow, contain limited samples, and exhibit inconsistent annotations or variable image quality; thereby, restricting their clinical reliability. This paper presents a comprehensive review and comparative analysis of fundus image datasets used in the management of DR. It introduces a taxonomy to organize the DR datasets and categorize them by size, accessibility, and annotation type (such as image‐level, lesion‐level, and multi‐disease). The study also evaluates their usability across key tasks, including binary classification, severity grading, lesion localization, and multi‐disease screening, based on an evaluation framework. A recently published dataset is presented as a case study to illustrate broader challenges in dataset curation and usage. Finally, this paper briefly discusses the recent data‐centric advancements, developments in deep learning models, and the impact of integrating clinical along with fundus images for automated screening. The review consolidates current knowledge while highlighting persistent gaps such as the lack of standardized lesion‐level annotations and longitudinal data. It also outlines recommendations for future dataset development to support clinically reliable and explainable solutions in DR screening.","PeriodicalId":501013,"journal":{"name":"WIREs Data Mining and Knowledge Discovery","volume":"47 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148633871","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Consumer Behavior Analysis in Digital Marketing Using AI and Big Data Analytics: A Narrative Review and Methodological Taxonomy","authors":"Leonidas Theodorakopoulos, Alexandra Theodoropoulou","doi":"10.1002/widm.70116","DOIUrl":"https://doi.org/10.1002/widm.70116","url":null,"abstract":"The rapid expansion of digital consumer data has challenged traditional approaches to understanding behavior in digital marketing. Existing reviews often focus on individual methods and give limited guidance on how analytical techniques compare or how they should be selected for specific marketing tasks. This article presents a structured narrative review of artificial intelligence and data‐driven techniques used in consumer behavior analysis, based on a curated body of peer‐reviewed studies. It organizes the field into key methodological families, including clustering, association rule mining, sentiment analysis, predictive modeling, time‐series forecasting, anomaly detection, path and sequence analysis, social network analysis, deep learning and multimodal analytics, generative AI and foundation models, agentic AI, causal analysis, and explainable AI. Building on this taxonomy, the paper offers a comparative synthesis of how each technique's conceptual assumptions, methodological reliability, data requirements, and typical outputs relate to its suitability for segmentation and profiling, personalization and targeting, fraud and risk detection, and engagement or journey optimization. The review also examines challenges related to data quality, integration, interpretability, scalability, and real‐time deployment, together with ethical and regulatory issues such as algorithmic bias, transparency, privacy, data sovereignty, and accountability. Finally, it outlines future research directions around self‐adaptive AI systems, agentic consumer interfaces, edge computing and localized intelligence, participatory data governance, quantum‐inspired analytics, and interdisciplinary methodological innovation. By linking analytical techniques to concrete marketing objectives under explicit ethical and regulatory constraints, the study provides researchers and practitioners with a structured basis for selecting, combining, and governing methods in consumer behavior analytics.","PeriodicalId":501013,"journal":{"name":"WIREs Data Mining and Knowledge Discovery","volume":"67 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148582243","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Exploring Large Language Model‐Based Intelligent Agents: Definitions, Methods, and Prospects","authors":"Yuheng Cheng, Ceyao Zhang, Zhengwen Zhang, Xiangrui Meng, Sirui Hong, Wenhao Li, Zihao Wang, Zekai Wang, Feng Yin, Junhua Zhao, Xiuqiang He","doi":"10.1002/widm.70111","DOIUrl":"https://doi.org/10.1002/widm.70111","url":null,"abstract":"The concept of the intelligent agent represents a long‐standing pursuit in artificial intelligence. Recent breakthroughs in large language models (LLMs) have catalyzed a paradigm shift, enabling the development of sophisticated agents that exhibit advanced reasoning, planning, and tool‐use capabilities across diverse domains. These LLM‐based agents, which leverage natural language as a universal interface for cognition and interaction, are rapidly advancing from theoretical constructs to practical applications, ranging from autonomous task assistants to complex multi‐agent simulations of social and economic systems. This paper provides an integrative survey of this burgeoning field. We first establish an organizing framework for understanding LLM‐based agents, systematically deconstructing both single‐agent and multi‐agent systems into their core components. We analyze the architectural principles and key mechanisms that underpin their intelligence, including planning paradigms, memory structures, and reflection‐based self‐improvement. We further investigate the dynamics of multi‐agent systems, exploring coordination strategies, communication protocols, and organizational structures. The paper also covers the crucial aspects of performance evaluation, highlighting influential benchmarks and identifying key challenges. Finally, we synthesize the current landscape to discuss the primary challenges, such as the intrinsic limitations of LLMs and the complexities of ensuring safety and alignment, and chart a course for future research directions, including the drive toward continual learning and enhanced multi‐modal capabilities.","PeriodicalId":501013,"journal":{"name":"WIREs Data Mining and Knowledge Discovery","volume":"26 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148582242","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Remote Sensing Techniques Shaping the Future of Wildfire Early Detection","authors":"Muneeb Raza, Xu‐Cheng Yin, Touqeer Abbas, Israr Ahmad, Rana Waqar Aslam, Muhammad Farukh Tanveer, Henrique Morimitsu","doi":"10.1002/widm.70094","DOIUrl":"https://doi.org/10.1002/widm.70094","url":null,"abstract":"The global wildfire crisis requires comprehensive early detection strategies to reduce environmental, economic, and societal impacts. This review systematically evaluates remote sensing (RS) technologies employed in early wildfire detection, classifying them into satellite systems, aerial platforms, and ground‐based sensors. Subsequently, it critically examines detection methodologies, which include image classification, object detection, semantic segmentation, and emergent approaches utilizing large language models (LLMs). Analysis reveals that deep learning (DL) significantly improves detection accuracy; however, its effectiveness is limited by resolution constraints, especially in the context of detecting smaller fires. As an advanced extension of the DL paradigm, LLMs built upon transformer‐based architectures complement conventional DL methods by enabling contextual reasoning over multimodal inputs, thereby bridging the gap between raw visual feature extraction and high‐level semantic interpretation. In this regard, LLMs demonstrate considerable potential in synthesizing complex, heterogeneous datasets, thereby enhancing the integration of diverse remote sensing data. Despite these advancements, substantial challenges remain, including limitations in spatial and temporal resolution, environmental interference, and complexities associated with data processing. To mitigate these obstacles, prospective research endeavors should prioritize strengthening algorithmic robustness, refining data fusion methodologies, and establishing real‐time monitoring systems. By integrating existing methodologies and outlining key challenges, this study aims to guide future academic inquiry in enhancing wildfire risk mitigation strategies and improving early detection frameworks. This article is categorized under: <jats:list list-type=\"simple\"> <jats:list-item> Fundamental Concepts of Data and Knowledge &gt; Big Data Mining </jats:list-item> <jats:list-item> Technologies &gt; Machine Learning </jats:list-item> <jats:list-item> Technologies &gt; Artificial Intelligence </jats:list-item> </jats:list>","PeriodicalId":501013,"journal":{"name":"WIREs Data Mining and Knowledge Discovery","volume":"40 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-07-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148564506","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Image Databases for Pollen Grain Classification: Availability, Quality, and Challenges","authors":"Heloise Acco Tives Bedin, Simone de Ramos, André Roberto Ortoncelli, Paula Fernandes Montanher, Patricia A. Jaques","doi":"10.1002/widm.70114","DOIUrl":"https://doi.org/10.1002/widm.70114","url":null,"abstract":"Pollen grain identification and categorization have applications in various fields, such as biology, medicine, and food safety. Traditionally, researchers manually perform pollen identification, a time‐consuming process prone to errors. With advancements in artificial intelligence and computer vision, there is growing interest in automating this task to improve efficiency and accuracy, making the availability of high‐quality labeled image databases essential. This study conducted a systematic literature review across six databases (ACM DL, DBLP, ScienceDirect, IEEE Xplore, Springer, and SOL) to map pollen image databases used over the past 10 years. From 1475 initially retrieved studies, rigorous inclusion and exclusion criteria reduced the sample to 73 relevant publications for final analysis. The study analyzed quantitative aspects, such as the number of species and image volume, and qualitative aspects, such as labeling methods and image processing techniques. The analysis identified 18 datasets and three crowdsourced databases, highlighting challenges such as the scarcity of available datasets, geographical concentration of samples, and the need for standardization in image annotation. The results emphasize the importance of expanding dataset diversity and improving image quality to enhance the application of machine learning models in pollen grain classification.","PeriodicalId":501013,"journal":{"name":"WIREs Data Mining and Knowledge Discovery","volume":"172 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-07-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148459813","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An Ontology Alignment: Systematic Literature Review of Datatypes, Methods, Evaluation Metrics, and Application Domain","authors":"S. Maria Senthamizh Selvi, V. Neelanarayanan, G. Y. Mythili","doi":"10.1002/widm.70112","DOIUrl":"https://doi.org/10.1002/widm.70112","url":null,"abstract":"Ontology alignment is the process of identifying equivalences between two ontologies to achieve semantic interoperability across systems. As research in this area continues to grow rapidly, it is essential to understand current trends and developments. This systematic literature review (SLR) is devoted to recent studies and analyzes key characteristics, including datasets, alignment methods, evaluation measures, and application areas. It is also bringing to the fore contemporary techniques, shortcomings, and gaps in research to be addressed in the future. The SLR was carried out through recent articles published in reputable journals. Primary studies were used to systematically gather features related to datasets, methodologies, approaches, and application domains engaged in the process of ontology alignment. The findings of this review indicate that new trends in ontology alignment primarily apply to machine learning‐based, automated, and hybrid approaches. Domain‐specific ontology datasets are commonly used, and OAEI continues to be a popular testing and evaluation benchmark. Nevertheless, major limitations persist around scalability, semantic heterogeneity, and the inability to support dynamic changes. Most assessment techniques focus narrowly on accuracy. Moreover, variations in reporting styles across different articles make comparison a hard task. Ontology alignment thus remains an evolving branch that requires further enhancement. Future research should consider a wider range of datasets spanning different domains, sizes, and formats. Comparison is expected to improve as researchers embrace standardized evaluation techniques.","PeriodicalId":501013,"journal":{"name":"WIREs Data Mining and Knowledge Discovery","volume":"324 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-07-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148460038","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}