NIMS: MagIC - Documentos de conferências internacionais
URI permanente para esta coleção:
Navegar
Entradas recentes
- The Hidden Side of Digital InclusionPublication . Macedo, Joana; Neves, Joana; Neves, Catarina; Oliveira, Tiago; Cruz-Jesus, Frederico; NOVA Information Management School (NOVA IMS); Information Management Research Center (MagIC) - NOVA Information Management SchoolDigital inclusion extends beyond access to platforms—it requires individuals to feel competent and autonomous in online spaces. While much research focuses on how social media affects mental health, this study examines the inverse: how poor mental health shapes experiences of digital participation. Drawing on self-determination theory, we analyze the effects of sleep quality, anxiety, and depression on perceived social media competence, autonomy, and relatedness. Results from a structural equation model show that anxiety and depression reduce competence and autonomy, while poor sleep undermines relatedness. Age moderates the link between relatedness and competence, with older users reporting lower perceived competence. These findings suggest that internal psychological states act as hidden barriers to digital inclusion, especially across age groups. The study contributes to conversations on identity and inclusion by highlighting how mental health influences one’s ability to engage meaningfully in social media environments.
- Exploring Elitism Strategies in Nested Tournament Selection for Multi-Objective Genetic ProgrammingPublication . Pereira, Filipa; Rebuli, Karina Brotto; Giacobini, Mario; Vanneschi, Leonardo; Information Management Research Center (MagIC) - NOVA Information Management School; NOVA Information Management School (NOVA IMS)Nested Tournament (NT) is a Multi-Objective (MO) selection method that enables fine-grained control of selection pressure through sequential single-objective tournaments. Although previously proposed, the impact of elitism strategies within NT remains largely unexplored. This study systematically investigates multiple elitism mechanisms for NT within tree-based MO Genetic Programming (MOGP), including NSGA-II population replacement, crowding distance, first-objective, and a novel ideal-point strategy, comparing them against non-elitist NT and standard NSGA-II. Experiments are conducted with up to five objectives for the accuracy-complexity trade-off and show that elitism design critically influences stability, efficiency, and semantic diversity. Notably, simpler NT-specific elitism strategies achieve comparable performance to NSGA-II at lower computational cost while better preserving semantic diversity. Overall, the findings highlight NT as an efficient MOGP alternative to selection methods based on Pareto Fronts.
- Designing LLM Agents for Output Checking in Research Data CentersPublication . Ashofteh, Afshin; Carvalho, Ricardo; Campos, Pedro; NOVA Information Management School (NOVA IMS); Information Management Research Center (MagIC) - NOVA Information Management SchoolResearch data centers increasingly face a tension between rapidly growing demand for microdata-driven research and strict confidentiality obligations. Output checking—reviewing tables, models, and descriptive statistics before release from safe centers-remains a key safeguard but is labor-intensive and difficult to scale. This paper presents a governed, semi-automated output-checking architecture that integrates three layers: (i) deterministic rule-of-thumb disclosure checks, (ii) a machine-learning classifier trained on historical release decisions to support triage, and (iii) a large language model (LLM) agent that performs principles-based synthesis and produces structured explanations for auditors and researchers. Using an archival corpus of output requests and released/blocked outputs from a national statistical office, we build a metadata and standardization pipeline for heterogeneous datasets and outputs, engineer disclosure-relevant features, and evaluate a prototype that routes low-risk cases quickly while escalating ambiguous cases for human control. We evaluate the architecture retrospectively on historical output-checking records and distinguish three validation targets: deterministic detection of obvious disclosure risks, ML prediction of institutional outcomes, and LLM-based evidence synthesis for audit support. The contribution is a governance-oriented architecture and empirical prototype for routing and explanation, not an autonomous release mechanism: all unsupported, ambiguous, or high-risk cases remain subject to professional output-checker review.
- High-Frequency Early Warning of Systemic Financial Stress in Europe Using Financial and Non-Financial Data with Machine LearningPublication . Diachkov, Dmytro; Ashofteh, Afshin; NOVA Information Management School (NOVA IMS); Information Management Research Center (MagIC) - NOVA Information Management SchoolEarly detection of systemic financial stress is challenging in fast-moving, nonlinear environments. Traditional early warning systems rely on low-frequency indicators and linear models, limiting their real-time relevance. This paper develops a high-frequency machine-learning framework to monitor systemic financial stress in Europe using daily financial, banking, macro-financial, and sentiment indicators. Diverse data sources are integrated in a strictly time-ordered pipeline to avoid look-ahead bias, with systemic financial stress states defined by threshold exceedances of the ECB's Composite Indicator of Systemic Stress. Feature selection and evaluation are conducted using stability-based regularization, time-series cross-validation, and rare-event metrics. Models combining financial and sentiment indicators outperform linear benchmarks, particularly during periods of intensifying stress. The results suggest that high-frequency machine-learning models can improve the timeliness of systemic stress monitoring and complement existing macroprudential tools.
- GenAI Usage in Higher EducationPublication . Aparicio, Manuela; Schroll, Kasper; Bernardo, Diogo Filipe Dinis; NOVA Information Management School (NOVA IMS); Information Management Research Center (MagIC) - NOVA Information Management SchoolGenerative AI (GenAI) is rapidly reshaping higher education, yet little is known about what drives students to continue using these tools over time. This study explores the evolution of students’ perceptions through two survey-based structural models (2023 and 2025), analyzed using SEM/PLS. Results reveal a striking shift: while students initially viewed GenAI with skepticism, by 2025 they reported satisfaction and perceived academic benefits. Early adoption was influenced by educational level, effort, and performance expectancy, whereas long-term use depended on system quality, satisfaction, individual impact, and ethical concerns. Trust, though not central at first, later influenced students’ sense of fairness and ethics. These findings highlight the importance of aligning GenAI tools with students’ expectations, values, and learning needs to ensure responsible and sustained use. This research offers valuable guidance for institutions aiming to integrate GenAI effectively and ethically into academic environments.
- Institutional design thinking for entrepreneurial governancePublication . Victorino, Guilherme; Martins, João; Neto, Miguel de Castro; Information Management Research Center (MagIC) - NOVA Information Management School; NOVA Information Management School (NOVA IMS)Many higher education institutions (HEIs) have come to believe that successful graduates represent the pinnacle of success. The institution’s focus on graduate employability may inadvertently suppress entrepreneurial behavior in its graduates. The purpose of this paper is to investigate this paradox and argue that the institutional successes in graduate employability may inhibit entrepreneurial behavior in graduates. This paper focus on one leading European data science and information management school showcasing exceptional levels of graduate employment and partnerships with both private and public sector organizations. In spite of these very positive conditions, however, there appear to be very few start-ups formed by graduates from the school. It appears that the problem lies in the design of the institutional governance structure of the school rather than a lack of entrepreneurial initiatives. When the curriculum, partnerships, incentives, and performance metrics for students are all primarily focused on preparing students to supply talent to existing organizations, then student expectations and career development are shaped toward employment rather than starting ventures. In order to help resolve this challenge, the paper presents a design-thinking based framework for designing entrepreneurial governance that includes three phases: Inspire, Ideate, and Implement. The first phase, Inspire, involves using self-assessment and stakeholder analysis to identify the contradictions between the school's rhetorical commitments to innovation and its actual entrepreneurial outcomes. The second phase, Ideate, identifies how governance structures, curricula, and incentives can be re-designed to encourage the formation of ventures while still providing students with the opportunity to pursue employment in traditional ways. The third phase, Implement, involves transforming entrepreneurship into an embedded part of the school's operations by establishing venture pipelines, creating new performance metrics, and increasing ties to entrepreneurial ecosystems. The proposed framework represents entrepreneurship in universities as a systemic design challenge rather than solely as an educational intervention. The paper also asserts that entrepreneurship cannot be successfully taught without corresponding systemic entrepreneurial governance structures. Therefore, by synchronizing strategy, incentives, partnerships, and learning environments, universities can convert successful graduate employment outcomes into increased entrepreneurial activity in the regions in which they are located and contribute more effectively to regional innovation ecosystems.
- Machine Learning Prediction Approaches to Bike-Sharing SystemsPublication . Oliveira, Rita; Jardim, Bruno; Albuquerque, Vitoria; Neto, Miguel de Castro; Information Management Research Center (MagIC) - NOVA Information Management School; NOVA Information Management School (NOVA IMS)Cities are adopting innovative mobility strategies to tackle urban challenges and mitigate the effects of climate change, emphasizing the integration of shared modes, such as bike-sharing systems. This study introduces a systematic literature review focusing on predictive machine learning (ML) methods’ contributions to bike-sharing systems (BSS) to promote effective implementation, sustainability, and user adoption of such systems. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) method was employed resulting in an analysis of 27 papers published between 2018 and 2023, identifying the most frequently used ML techniques in this domain and offering an outline for future research. This study identified RF, GBT, and XGBoost models as the most used predictive modeling methods due to their robust performance in predicting station-level BSS supply and demand. The review also highlights the importance of incorporating diverse features such as weather, time, spatial factors, and socio-economic data, which significantly enhance model accuracy. This study provides a state-of-the-art about ML techniques utilized for predicting station-level BSS supply and demand to guide future research directions.
- Predicting Station Occupancy of Bike-Sharing System During EventsPublication . Oliveira, Rita; Jardim, Bruno; Albuquerque, Vitoria; Neto, Miguel de Castro; Information Management Research Center (MagIC) - NOVA Information Management School; NOVA Information Management School (NOVA IMS)Technological advancements have enabled the large-scale collection of bike-sharing systems (BSS), providing valuable opportunities for city planners and operators to optimize services and improve decision-making. As a soft mobility mode, BSS plays a critical role in promoting sustainable and reliable urban transportation. However, research remains limited regarding how such data can be used to evaluate the impact of city events on BSS usage and to enhance station occupancy prediction. To address this gap, this study uses Lisbon, Portugal, as a case study to investigate the influence of events on BSS usage, focusing on hourly station occupancy rates and, for the first time in the city, integrating event data into BSS prediction models. By integrating weather conditions and event data, we apply state-of-the-art machine learning algorithms, such as Random Forest, Gradient Boosting Tree, and Extreme Gradient Boosting, to predict occupancy levels. The goal is to evaluate model performance, analyze prediction error patterns, and contribute to more efficient resource allocation and management of BSS during urban events. Results show that an XGBoost model enriched with an occupancy-change feature and historical demand data delivers the most accurate station-level occupancy predictions, revealing distinct usage patterns between sports and music events and across station locations, and provides actionable insights to support targeted, time-sensitive bicycle rebalancing during high-impact event days. The study provides a benchmark methodology that can be used in other urban contexts, enabling smart cities’ advancements.
- Centro Nacional de Dados Oceanográficos (NODC-PT)Publication . Dias, Telmo; Fortes, Isabel; Alves, Margarida; Esteves, Rita; Dias, Elisabete; Barata, Paulo; Costa, Ana Cristina; Baptista, Márcia L.; NOVA Information Management School (NOVA IMS); Information Management Research Center (MagIC) - NOVA Information Management SchoolThe Portuguese National Oceanographic Data Centre (NODC-PT) was established to improve the management and FAIRness of marine data produced by Portuguese institutions. Developed in coordination with the Portuguese Committee for the Intergovernmental Oceanographic Commission (IOC), NODC-PT adopts a federated architecture that promotes data sharing while preserving the autonomy of data providers. The infrastructure supports metadata harmonization, interoperable access services, and data stewardship practices aligned with FAIR principles (Findable, Accessible, Interoperable and Reusable). Integrated within the IOC International Oceanographic Data and Information Exchange (IODE) network, NODC-PT contributes to national and international initiatives related to ocean digitalization and the United Nations Decade of Ocean Science for Sustainable Development. This paper presents the main architectural principles, integration models, and governance approaches adopted by NODC-PT, highlighting its contribution to marine data reuse, marine research strengthening, and long-term preservation of oceanographic information.
- From Six to TwoPublication . Rosenfeld, Liah; Vanneschi, Leonardo; Information Management Research Center (MagIC) - NOVA Information Management School; NOVA Information Management School (NOVA IMS)Geometric Semantic Genetic Programming (GSGP) is an extension of Genetic Programming (GP) that captured the interest of researchers because of its ability to induce a unimodal error surface for any supervised learning problem. Although still a recent development, the Semantic Learning with Inflate and Deflate Mutations (SLIM-GSGP) extension of GSGP has already attracted significant attention due to its novel ability to generate offspring that are smaller than their parents, effectively addressing the problem of steady model growth in GSGP. This paper presents SLIM-DUO, an extension of SLIM-GSGP that integrates the six existing SLIMGSGP variants into solely two unified formulations: DUO-MUL and DUO-SUM. This integration streamlines benchmarking and hyperparameter exploration while aiming to retain comparable predictive performance at approximately one third of the computational cost. Across five test problems, SLIM-DUO achieves predictive performance comparable to that of SLIM-GSGP, with model size differences that remain within previously observed dataset-dependent variability among SLIM variants. Overall, SLIM-DUO substantially reduces the computational effort required during both the configuration and benchmarking phases, highlighting a favorable balance between model size and computational complexity and preserving solution quality.