Over the past few decades, scholars' dedication to addressing global challenges, including social, economic, management, and environmental issues, has grown. This has led to a surge in scientific publications, especially review studies that aim to provide meaningful theoretical, practical, and policy-level implications for addressing these challenges. Despite this development, the rapid growth has raised concerns among researchers about the originality, transparency, reproducibility, reliability, and validity of the conclusions of these review studies. To help address this issue, particularly in the social sciences, this study contributes to the extant literature by presenting a detailed methodological framework that builds on current approaches and guides the conduct of high-quality systematic literature reviews and bibliometric analyses. To help scholars appreciate the differences among review types, we provide an overview of various review articles—including narrative reviews, systematic literature reviews, scoping reviews, integrative reviews, bibliometric analyses, meta-analyses, scientometric analyses, informetric analyses, webometrics, and patent analyses—and their characteristics to enhance understanding. We then demonstrate, through a practical example, how to integrate emerging digital innovation tools, such as VOSviewer and the R package Bibliometrix, to extract, process, and manage large datasets that can complement systematic literature reviews. More importantly, we illustrate, with an example, how scholars can leverage the specialized functions of bibliometric tools, such as co-word analysis, bibliographic coupling, word clouding, thematic mapping, and co-authorship (social collaboration) analysis, to identify research gaps and themes while adopting some of the best practices that can contribute to knowledge advancement within the social science field. For standard and high-quality reviews, we argue that this study will broaden scholars' and information managers' perspectives and promote a paradigm shift, guiding the growth and development of the academic landscape. It also serves as a resource for early-career researchers, seasoned scholars, journal reviewers, and policymakers, especially those new to this type of research.
Image captioning, the task of generating textual descriptions for images, has advanced significantly with datasets like MSCOCO and models like Transformers. However, these developments have largely benefited high-resource languages like English, leaving low-resource languages such as Tamil underrepresented. Tamil, a Dravidian language with over 80 million speakers, faces challenges in image captioning due to the lack of resources that capture its unique linguistic and cultural nuances. Existing datasets often rely on direct translations, which fail to address semantic misalignments and cultural specificity, limiting their effectiveness for Tamil. Motivated by this gap, this paper introduces TamilCOCO, a novel dataset comprising 63,062 images with 305,340 culturally adapted English-Tamil caption pairs curated from MSCOCO with culturally adapted Tamil captions. TamilCOCO uses a semi-automated annotation framework that integrates multilingual models, a Cultural Adaptation Module (CAM), and iterative community validation, ensuring both semantic accuracy and cultural fidelity. The dataset was evaluated using metrics-BLEU (0.68), METEOR (0.83), CIDEr (1.48), and SPICE (0.53)-alongside a Cultural Relevance Score (CRS) averaging 0.86, highlighting its cultural alignment. Fine-tuning baseline models, including MuRAG, achieved significant improvements, with CRS reaching 0.88. TamilCOCO not only establishes a benchmark for low-resource image captioning but also provides a scalable framework for adapting datasets to other culturally rich, low-resource languages.
The widespread use of data presents significant managerial challenges, particularly when this information is legally protected by privacy regulations. When managing a large amount of data with dispersed rightsholders, obtaining their legal consent incurs high transaction costs. Collective governance models like data trust offer a potential solution. However, traditional trust objects differ from personal data, as it involves personal rights that are essentially inalienable in general. Therefore, negotiating personal rights through an economic scheme involves complex incentives and cost-benefit trade-off problems, presenting economic inefficiencies. Trust and privacy regulations, while important, are limited frameworks in this context, but the copyright system can complement and reconcile the legal nature of the collectively managed rights over intangible assets. This cross-disciplinary research proposes migrating data trust agreements to data collective management entities, combining legal-dogmatic and economic approaches with systems thinking methodologies. This proposal offers an integrated and diverse perspective for ongoing academic discussions and for future legislation and public policies in the field, potentially leading to more efficient and sustainable data governance and information management.
In the context of the accelerating digital transformation, Information Technology (IT) capabilities have become a core driver for firms to achieve innovation performance. Based on the TOE (Technology- Organization-Environment) theoretical framework, this study employs bibliometric and content analysis methods to systematically review and empirically integrate 298 high-quality English articles from the Web of Science Core Collection published between 2016 and 2025. The aim is to uncover the multi-dimensional collaborative mechanisms through which IT capabilities impact firm innovation performance. The research finds that technological IT resource infrastructure capabilities and emerging digital technology capabilities form the critical technological foundation driving corporate data insights and technology conversion. Organizationallevel IT resource integration capabilities and IT human resource capability act as essential intermediaries linking technology deployment to performance realization. At the same time, market environmental dynamism, industry characteristics, and government policy significantly moderate the capability conversion paths. Furthermore, the study proposes a closed-loop model of "system embedding-capability integration- performance feedback", highlighting the path dependence and conditional adaptability of IT capabilities in multi-level interactions and dynamic environments. This research not only enriches the understanding of the role mechanisms of IT capabilities in the field of information systems but also provides theoretical foundations and practical insights for firms in formulating digital transformation strategies and achieving the system conversion of technological value.
Analyzing customer perceptions has become increasingly important in the automotive industry as it provides actionable insights into consumer satisfaction, preferences, and areas requiring improvement. This study proposes a novel perception analysis framework for automotive reviews using advanced Large Language Models (LLMs), including BERT, FLAN-XXL, and Mistral 7B, leveraging zero-shot learning to categorize reviews without task-specific training data. The framework follows a two-stage evaluation process, beginning with zero-shot perception classification and followed by a detailed topic-wise perception analysis. Model performance was evaluated using accuracy, precision, recall, and F1-score across reviews from five major automotive brands—Toyota, Kia, Honda, Nissan, and Hyundai. While BERT and FLAN-XXL achieved high accuracy in certain cases, they exhibited limitations in balancing precision and recall, particularly for nuanced perceptions. In contrast, Mistral 7B consistently demonstrated superior F1-score performance and was therefore selected for in-depth topic-wise analysis across key automotive aspects, including comfort, safety, performance, design, and price. The results indicate that Mistral 7B excels in objective categories such as performance and price, while performance declines for more subjective topics such as comfort and safety. Overall, the findings highlight the effectiveness of zero-shot LLM-based perception analysis for automating large-scale customer feedback analysis in the automotive domain. Future work will focus on domain-specific fine-tuning and the integration of multimodal information to further improve performance, particularly in subjective perception categories, enabling more accurate and data-driven decision-making in the automotive industry.
