Data and Metadata (Journal)
Not a member yet
727 research outputs found
Sort by
Impact of data protection legislation on the digitalization of small and medium-sized enterprise
Full text linkIntroduction: Decision-making in small and medium-sized enterprises (SMEs) relies heavily on the proper management of data. Objective: The objective of this research was to describe the scientific output on data protection legislation in the digitization of small and medium-sized enterprises, which is required by these enterprises for decision-making. Methods: To this end, the Scimago portal was used as a source for analyzing scientific output, and a review of the distribution by quartiles was carried out based on bibliometric indicators such as the H index, the impact factor of journals, the number of documents published, the average number of citations per document, international scientific collaboration, and citations from public funding agencies. Likewise, the number of articles cited in the most relevant journals in each quartile was analyzed in order to assess the relevance of scientific innovation in the formulation of adequate data protection legislation in the process of digitizing SMEs. Results: The findings reveal that most scientific output is concentrated in journals belonging to countries with a high level of technological development. The areas with the highest representation are threat detection and data protection. Conclusions: A low level of dissemination of research related to the development of specific legislative frameworks for data protection in the context of the digitization of small and medium-sized enterprises was identified.
Exploring Deep Learning Approaches for Multimodal Breast Cancer Dataset Classification and Detection
Full text linkIntroduction; Globally, we need advanced testing to detect breast cancer early. New breast cancer diagnosis methods using mixed datasets and deep learning promise improved accuracy. Objective; These sets, which comprise several imaging modalities, show tumor characteristics well. VGG16, AlexNet, and ResNet50 are effective deep learning models in many domains, yet their breast cancer diagnosis performance is unclear. Method; This paper examines these patterns\u27 benefits, downsides, and research gaps. We also provide two novel approaches, Attention-based Multimodal Fusion (AMF) and Improved Generative Adversarial Augmentation (GAA), to improve deep learning models on breast cancer datasets. Result; The findings highlight the potential of machine learning to show tumor characteristics well. Conclusion; We prove that our breast cancer screening technologies are the most accurate and dependable via extensive testing
Detection and Prediction of Financial Fraud Using Deep Learning Methods: A case of the Companies Listed in the Amman Stock Exchange
Full text linkIntroduction: The study examined the ongoing issue of identifying financial fraud in emerging economies, concentrating on companies listed on the Amman Stock Exchange (ASE).Methods: A panel of 176 ASE-listed enterprises was studied from 2011 to 2021. Starting with a preliminary analysis of Beneish M-Score constituents and associated metrics, a supervised neural network (FNN) had been trained, and an ordinary least-squares (OLS) analysis was computed. The performance study was executed using reliability, recall, reliability, F1-score, and ROC-AUC.Results: The FNN achieved an accurate identification rate of 0.9844 with a recall of 1.0, indicating it accurately identified all fraudulent transactions in the experimental dataset. The ROC-AUC was 0.97. The OLS model, albeit less precise, demonstrated statistically significant correlations—particularly for GMI, SGAI, and LVGI—with the Beneish M-Score, thereby providing interpretable risk indicators.Conclusions: The study revealed that deep learning, namely a feedforward neural network (FNN), surpassed a traditional ordinary least squares (OLS) method in detecting fraud among ASE enterprises, whereas OLS offered contextual information about the factors associated with fraud. An integrated analytical framework was proposed to assist regulators and investors in achieving improved transparency and early warning in the Jordanian market
What Do Scopus Index Keywords Reveal About Educational Data Mining Research? A Bibliometric Analysis (2014-2024)
Full text linkIntroduction: Educational Data Mining (EDM) has emerged as a pivotal interdisciplinary field addressing the increasing demand for data-driven educational enhancement. However, a comprehensive understanding of its developmental trajectory is hindered by fragmented literature reviews and a lack of longitudinal analysis spanning critical technological and educational transformations.Objective: This study investigates the evolution of EDM research over the transformative decade from 2014 to 2024 through systematic bibliometric analysis, aiming to identify growth patterns, thematic developments, and methodological innovations.Method: We conducted an extensive analysis of 436 peer-reviewed publications indexed in Scopus, employing rigorous keyword analysis, mathematical modeling of research trends, and systematic thematic classification to examine temporal evolution patterns. The methodology utilized PRISMA-guided selection procedures, standardized keyword extraction and normalization, and quantitative measures including growth ratios, Shannon diversity indices, and thematic strength calculations.Results: Our analysis reveals remarkable research growth, with a 777.8 % increase in publication output, representing a compound annual growth rate of 24.3 %. The findings document a significant paradigmatic shift from descriptive analytics toward predictive methodologies, evidenced by a 215-fold growth in Machine Learning and AI themes and the complete emergence of deep learning applications. Thematic evolution analysis identified 47.3 % of recent keywords as entirely new terms, indicating substantial conceptual expansion.Conclusions: The research demonstrates EDM\u27s transition from foundational exploration (2014-2017) through rapid expansion (2018-2020) to sophisticated maturation (2021-2024), characterized by methodological pluralism and the integration of advanced computational techniques.
Simulated Annealing-Based Optimization of IEEE-33 Radial Distribution Networks with Integrated Auxiliary PV Sources Using PyPower
Full text linkThe integration of distributed photovoltaic (PV) generation in radial distribution networks has been consolidated as a key strategy to reduce technical losses and improve voltage profiles. However, the optimal placement and sizing of these sources remain a challenge due to the nonlinear and multimodal nature of the problem. In this work, an approach based on simulated annealing with adaptive restarts, implemented in PyPower, is proposed to determine the optimal location of PV units in the IEEE 33-bus test system. The methodology considers the minimization of active power losses as the objective function, subject to operational constraints and voltage limits.The results show that the proposed strategy achieves a reduction in losses of up to 52% compared to the base scenario, in addition to improving minimum voltage profiles to values close to 0.98 pu. The comparison with non-optimized scenarios highlights the effectiveness of the method in balancing energy efficiency and quality of service. This study contributes to the literature by demonstrating the applicability of lightweight metaheuristics in distribution network planning problems and lays the groundwork for future research integrating storage and dynamic load and generation scenarios
Gamification and Formative Feedback with Kahoot! in Programming Fundamentals: A TAM–SDT–UCD Approach with Mechatronics Engineering Students
Full text linkImmediate and meaningful feedback is a key element in improving academic performance in computational foundational disciplines. This study aimed to evaluate the impact of gamified strategies in the feedback process of the Programming Fundamentals course. The intervention was conducted with a population of 31 first-level students from the Mechatronics Engineering program at Universidad Técnica del Norte, of which 15 students voluntarily participated from the same course. Activities were designed using the Kahoot! platform, focused on reinforcing theoretical and practical content related to the course syllabus and the programming languages Java and Python. The adopted methodology was applied in nature, with a descriptive approach and exploratory scope. The results showed increased active participation, improved retention of key concepts, and a more positive attitude towards programming learning. In addition, recurring error patterns were identified, which served as input for tailored feedback. The use of playful dynamics fostered a collaborative learning environment, contributing to the development of cognitive and metacognitive skills. It was concluded that the integration of educational technologies with gamification elements enhances the effectiveness of feedback processes, increases students\u27 intrinsic motivation, and enables timely identification of knowledge gaps. This approach is especially valuable in programming education, where active engagement and continuous practice are critical to academic success
AI in Scientific Research: Automation, Original Discovery, and Ethical Collaboration
Full text linkGenerative artificial intelligence (AI) is dramatically changing scientific research processes because of its ability to contribute to the various phases of the scientific method, allowing researchers to accelerate the scientific production process. However, the obstacles and restrictions posed by the use of AI are significant due to biases related to the ethical use of models and frameworks that foster a sense of responsibility and ethical principles in the application of AI. The aim of this study is to identify the main research trends related to the use and application of AI in scientific production; accordingly, this article presents a systematic literature review using the Preferred Reporting Items for Systematic Reviews and Me-ta-Analyses model (PRISMA). A total of 4,334 records were obtained from the Scopus and Web of Science databases and thoroughly reviewed, selecting 233 for qualitative and quantitative analyses. Four major trends in the use of AI in scientific research were identified, associated with the conceptual cycle of the scientific method: 1) scientific discovery, 2) data management, 3) academic writing, and 4) scientific communication. The ethical considerations highlighted in the literature advocate for the establishment of clear policies and frameworks within academic institutions that enable researchers to use the tools effectively throughout educational processes. Rather than focusing on debates about the use of AI, this approach promotes an ethical and efficient integration of AI into scientific.
Data-Driven Evaluation of Employee Performance as a Mediator between Technology, Work Culture, and Service Quality
Full text linkIntroduction: technology adoption, work culture, and servant leadership are critical organizational factors influencing service quality in public sector institutions. Understanding their impact requires a data-driven approach to evaluate how these variables interact through employee performance. This study aims to provide a quantitative assessment of the relationships between organizational factors, employee performance, and service quality within Indonesia’s national social security agency, highlighting the mediating role of employee performance.
Method: data were collected through a structured questionnaire administered to employees in April 2025. Using Structural Equation Modeling (SEM), the study analyzed the direct and indirect effects of technology adoption, work culture, and servant leadership on service quality via employee performance. Metadata-driven techniques were applied to organize survey responses and extract insights regarding the relative influence of each factor.
Results: the analysis reveals that technology adoption and work culture significantly enhance employee performance, which in turn positively mediates their effect on service quality. Servant leadership, however, does not demonstrate a direct influence on service quality, suggesting its role may operate through indirect or contextual mechanisms. Quantitative metrics indicate that employee performance serves as a key pathway linking organizational resources to service outcomes.
Conclusions: This data-driven assessment highlights the critical role of employee performance in translating organizational initiatives into improved service quality. The findings underscore the importance of fostering digital transformation and a strong work culture while integrating leadership strategies that indirectly support performance. By framing the study within a quantitative and metadata-informed perspective, the results provide actionable insights for policymakers and managers seeking to optimize service delivery in public sector organizations
Cybersecurity Challenges in Multimodal Medical Data: A Critical Review with a Focus on Diabetic Retinopathy Screening Systems
Full text linkIntroduction: this critical narrative review examined cybersecurity challenges in multimodal diabetic retinopathy (DR) screening systems, addressing the convergence of diverse data types within complex regulatory frameworks. With 537 million diabetics at risk globally and healthcare cyber incidents increasing by 45 % in 2023, the study investigated security vulnerabilities arising from integrating high-resolution imaging with clinical parameters.Methods: the review employed an iterative search strategy across PubMed/MEDLINE, IEEE Xplore, Scopus, ACM Digital Library, and arXiv. From 487 initially identified publications, structured extraction and full-text review yielded 50 high-quality sources. The analysis synthesized findings through complexity theory, developing the novel Diabetic Retinopathy Security Complexity Index (DRSCI) to quantify multiplicative security challenges.Results: the DRSCI revealed that 73 % of international collaborative screening programs exceeded manageable complexity thresholds (>1000), corresponding with vulnerability assessments showing 56 % of medical device vulnerabilities classified as critical or high-severity. The review identified critical gaps between theoretical security models and operational realities, particularly in multimodal data integration across jurisdictions. Current ISO 27799:2016 standards proved inadequate for addressing high-volume imaging data challenges.Conclusions: the multimodal nature of modern DR screening created vulnerability surfaces transcending traditional security paradigms. The DRSCI framework transformed abstract risk assessments into actionable metrics, enabling evidence-based security investment decisions. Immediate priorities included developing quantum-resistant algorithms, implementing federated learning frameworks, and establishing comprehensive multimodal security standards before projected quantum computing threats materialize by 2030
Data-Driven Redefinition of Radicalism-Terrorism: Understanding the Continuum of Intolerance and Extremism through Indonesian Experience
Full text linkIntroduction: existing definitions of intolerance, radicalism, extremism, and terrorism (IRET) in the global literature—such as Moghaddam’s Staircase to Terrorism, McCauley & Moskalenko’s Two-Pyramids Model, and Sageman’s social network approach—have shaped the theoretical foundation of radicalization studies. However, the Indonesian context presents distinctive dynamics that are not fully captured by these frameworks, including rapid radicalization through digital media, institutionalized intolerance within local policies, women’s roles in online propaganda, and the rise of hybrid terrorism that combines physical and cyber dimensions.Objective: this study aims to provide a data-driven redefinition of intolerance, radicalism, extremism, and terrorism (IRET) by integrating Indonesia’s contextual realities with established global theoretical models.Method: A qualitative–quantitative mixed approach was employed. The study critically engaged with existing theories, analyzed metadata and case data from terrorism management in Indonesia, and reviewed relevant scholarly contributions. Secondary datasets on terrorism cases, policy documents, and digital propaganda activities were systematically examined, while recent literature provided comparative perspectives to validate the proposed conceptual model.Results: findings demonstrate that traditional models of radicalization require adaptation to address emerging trends within the Indonesian context. Quantitative results indicate a significant correlation between social media exposure and the early stages of radicalization, highlighting the role of online networks in shaping extremist attitudes. Additionally, gender-based digital propaganda and hybrid forms of terrorism—combining physical and cyber elements—emerge as critical dimensions influencing radical behavior.Conclusion: the proposed data-driven redefinition of IRET incorporates these contemporary dynamics, offering a more comprehensive understanding of radicalization and terrorism in Indonesia. This framework enhances counterterrorism discourse by connecting context-specific insights with global theoretical debates. Policy implications include the need for integrated monitoring of digital radicalization, adaptive legal frameworks, and inclusive community-based prevention strategies