Ciências e Tecnologia | Artigos em revistas internacionais / Papers in international journals
Permanent URI for this collection
Browse
Recent Submissions
- Assessment of recreational boat anglers’ practices and perceptions in the Faial-Pico Channel marine protected area, AzoresPublication . Silva, Cliff; Seixas, SóniaThe Faial-Pico Channel Marine Protected Area (MPA), located in the Azores, Portugal, encompasses approximately 240 km² and is divided into two main zones: Faial and Pico. Unlike the broader regulations across mainland Portugal, the unique licensing requirements in the Azores only mandate boat owners to obtain licenses. This study aims to fill the existing research gap on recreational boat anglers in the Faial-Pico Channel MPA by employing face-to-face survey interviews with local boat anglers to understand their fishing practices, target species, and perceptions regarding marine resource management. Results from experienced recreational boat anglers indicate a predominant fishing duration of 11 to 30 days per year, with fishing efforts primarily targeting species such as forkbeard and blackspot seabream. The estimated annual capture is enormous, comprising a significant portion of the total fishing yield on Faial Island. Most boat anglers support specific conservation measures such as minimum length regulations and closed seasons, though there are varied opinions on the broader impacts of recreational fishing. Boat anglers largely attribute the decline in certain fish stocks to commercial overfishing. While the existing restrictions in Monte da Guia are accepted, proposals for additional MPAs face resistance. This study underscores the need for more comprehensive management strategies, such as improving communication between stakeholders and employing a licensing system to regulate fishing activities better and conserve marine biodiversity in the MPA.
- A machine learning framework for uplift modeling through customer segmentationPublication . Pinheiro, Paulo; Cavique, LuísIn uplift modeling, the goal is to identify high-value customers based on persuadable customers, those who make a purchase only if contacted. To achieve this, uplift modeling combines machine learning techniques with causal inference, allowing businesses to refine their customer targeting strategies and focus efforts where they are most profitable. This study proposes a practical and reproducible two-phase procedure for identifying highvalue customers. In the first phase, customers are segmented using decision trees, which offer a transparent and data-driven approach to grouping individuals with similar characteristics. This segmentation lays the groundwork for a meaningful interpretation of customer behavior. In the second phase, uplift is calculated for each customer segment by comparing the outcomes of the treatment and control groups. This enables the identification of customer groups with the highest uplift. A real-world use case further illustrates the value and applicability of the proposed method. To validate model performance, the procedure employs established metrics such as the Qini index and Cohen’s kappa, which provide insights into both the effectiveness and reliability of the uplift estimates. This work presents a decoupled procedure for uplift modeling that leverages well-established libraries, fostering transparency and a clear understanding of the analytical process. A key contribution to uplift modeling and causal inference is the use of decision trees for stratification, which enables the creation of meaningful segments and their evaluation through the average treatment effect. By integrating theory with practical implementation, this work offers a comprehensive framework for uplift modeling that bridges academic rigor and business usability.
- The reproductive ecology of Loligo squids (Cephalopoda: Myopsida) in Lusitanian zoogeographical provincePublication . Laptikhovsky, Vladimir; Moreno, Ana; Oesterwind, Daniel; Perales-Raya, C.; Pierce, Graham; Robin, J.-P.; Sobrino, Ignacio; Valls, M.; Villanueva, R.; Allen, A.; Abad, E.; Bello, G.; Bobowski, B.; Fernández-Álvarez, F.; Cabanellas-Reboredo, M.; González, Á.; Hernández-Urcera, J.; Krstulović-Šifner, S.; Lefkaditou, E.; MacLeod, E.; Valeiras , J.; Onsoy, B.; Pereira, João; Salman, A.; Seixas, Sónia; Power, A. M.Analysis of the temporal and spatial distribution of 686 records of egg masses and egg mass groups from (i) recreational divers and posted in the various public media, (ii) scientific survey data collected during research programs and (iii) publications in peer-reviewed literature plus published and unpublished information on the seasonality of the occurrence of mature females, demonstrated that the reproduction of Loligo squids occurs throughout the Lusitanian zoogeographical province all year round with varying seasonal peaks. This may be due to high phenotypic plasticity, with the life cycle adapting to local conditions, or the existence of small discrete stock units related to individual water features. In warmer and relatively productive waters the spawning peak is more extended, and percentage of mature females in catches is lower In the Mediterranean the peak of occur rence of both mature females and spawned eggs gradually shifts to earlier dates from the west to the east, which is consistent with respective changes of the earlier peak of productivity. There is little or no gap between peak of occurrence of mature females and peak of the egg mass records. Spawning grounds of L. vulgaris extend in the Mediterranean area much deeper than thought before and they extend from 3 m to 550 m. Egg masses were reported by recreational scuba divers from deeper locations in the central Mediterranean than in other areas. Loligo forbesii egg masses were found between 170 and 720 m.
- Mitigating false negatives in imbalanced datasets: an ensemble approachPublication . Cavique, Luís; Vasconcelos, MarceloImbalanced datasets present a challenge in machine learning, especially in binary classification scenarios where one class significantly outweighs the other. This imbalance often leads to models favoring the majority class, resulting in inadequate predictions for the minority class, specifically in false negatives. In response to this issue, this work introduces the MinFNR ensemble algorithm, designed to minimize False Negative Rates (FNR) in imbalanced datasets. The new approach strategically combines data-level, algorithmic-level, and hybrid-level approaches to enhance overall predictive capabilities while minimizing computational resources using the Set Covering Problem (SCP) formulation. Through a comprehensive evaluation of diverse datasets, MinFNR consistently outperforms individual algorithms, showing its potential for applications where the cost of false negatives is substantial, such as fraud detection and medical diagnosis. This work also contributes to ongoing efforts to improve the reliability and effectiveness of machine learning algorithms in real imbalanced scenarios.
- A data science maturity model applied to students' modelingPublication . Cavique, Luís; Pombalinho, Paulo; Correia, LuísMaturity models define a series of levels, each representing an increased complexity in information systems. Data Science appears in the Business Intelligence (BI) and Business Analytics (BA) literature. This work applies the _IABE maturity model, which includes two additional levels: Data Engineering (DE) at the bottom and Business Experimentation (BE) at the top. This study uses the _IABE model for students' modeling in the ModEst project. For this purpose, the Public Administration organism is the Directorate-General for Statistics of Education and Science (DGEEC) of the Portuguese Education Ministry. DGEEC provided vast data on two million students per year in the Portuguese school system, from pre-scholar to doctoral programs. This work presents the comprehensible _IABE maturity model to extract new knowledge from the DGEEC dataset. The method applied is _IABE, where after the DE level, wh-questions are formulated and answered with the most appropriate techniques at each maturity level. This work's novelty is applying the maturity model _IABE to a unique dataset for the first time. Wh-questions are stated at the BI level using data summarization; at the BA level, predictive models are performed, and counterfactual approaches are presented at the BE level.
- Implications of causality in artificial intelligencePublication . Cavique, LuísOver the last decade, investment in artificial intelligence (AI) has grown significantly, driven by technology companies and the demand for PhDs in AI. However, new challenges have emerged, such as the ‘black box’ and bias in AI models. Several approaches have been developed to reduce these problems. Responsible AI focuses on the ethical development of AI systems, considering social impact. Fair AI seeks to identify and correct algorithm biases, promoting equitable decisions. Explainable AI aims to create transparent models that allow users to interpret results. Finally, Causal AI emphasizes identifying cause-andeffect relationships and plays a crucial role in creating more robust and reliable systems, thereby promoting fairness and transparency in AI development. Responsible, Fair, and Explainable AI has several weaknesses. However, Causal AI is the approach with the slightest criticism, offering reassurance about the ethical development of AI.
- Assessment in collaborative learning: a mediation analysis approachPublication . Cavique, Luís; Ramos, Maria do RosárioIn collaborative learning, evaluating the process involves teamwork dynamics, and assessing the product focuses on the accuracy and quality of the final output. Assessment plays a crucial role, as it defines and measures the effectiveness of group activities to ensure that learning objectives are met. Mediation analysis is an important technique to better understand relationships between variables, specifically designed to test hypotheses about potential causal effects in various areas. However, many research initiatives have been discontinued prematurely due to the Baron-Kenny data restrictions. This research takes a case study of online learning from the Portuguese Open University to determine if and how group selection and interaction frequency affect individual assessment. The contribution lies in applying quantitative causal mediation analysis to collaborative learning assessment. The Lambda Mediation Ratio is proposed to enhance mediation analysis by enabling quick and flexible categorization into full, partial, or no mediation. Using Moodle platform logs and student outcomes, it was possible to find a significant influence of group dynamics on academic performance, highlighting the practical application of this improved methodology in an educational context. These findings reassure us of the relevance and applicability of this research in real-world educational settings.
- Putting fishing communities on the map in ICES ecoregionsPublication . Kraan, Marloes; Himes-Cornell, Amber; Pedreschi, Debbi; Motova, Arina; Hamon, Katell; Pita, Cristina; Ballesteros, Marta; Barz, Fanny; Fonseca, Tereza; García-De-Vinuesa, Alfredo; Guitierrez, Angel; Jackson, Emmet; Lam, Mimi; Norman, Karma; Seixas, Sónia; Steins, NathalieThis paper highlights the importance of identifying fishing communities for fisheries and ecosystem-based management, which often focuses on fleets and ecological impacts rather than on the communities where fishers live and land their catches. Fishing communities are key to understanding the broader impacts and benefits of fishing, as they support many livelihoods in fleet and trade-related activities. Recognizing these communities, allows for better data collection, analysis, and informed policy-making. ICES WGSOCIAL developed a method to identify fishing communities across ICES ecoregions, first applied in the Celtic Seas and North Sea ecosystem o verviews. These o verviews describe ecosystems, identify human pressures, and assess their impact. Using fishing ports as proxies, our method links socio-economic indicators (e.g. landings value) to communities. We identify limitations to our methods and explore the complexities of defining a ‘fishing community’ due to its dynamic, multidimensional nature. We discuss next steps for improving our mapping approach and deepening our understanding of the social, cultural, and economic value of fishing, and why these matter for applied marine science in support of policy and management.
- Walking to public transport: rethinking catchment areas considering topography and surrogate buffersPublication . Pais, Filipe; Sousa, Nuno; Monteiro, João Pedro Medina ; Rodrigues, João Coutinho; Jesus, Eduardo NatividadeService, or catchment areas of public transport stops are traditionally assessed using Euclidean or network distances, often neglecting other relevant factors such as topography. This study proposes a refined approach that integrates network-based accessibility with terrain variations and the effect they have on walking time and on the physical effort required for pedestrian movement. Using geographic information systems-based analysis that include walking time and walking energy cost models, the impact of topography on accessibility to public transport is evaluated in a case study of the hilly city of Coimbra, Portugal. Results show that, as compared to their flat counterparts, network distance-based service areas that consider hilliness, exhibit a decrease in accessibility of circa 10% in terms of area covered and population affected. These findings highlight the need for more realistic accessibility assessments to support more realistic and equitable public transport planning. Because extensive network datasets are not always available to decision-makers, this article also introduces the concept of surrogate buffers as a practical alternative for obtaining catchment areas, summarized by the “0.7/0.6R rule”.
- Determinants of fare evasion in urban bus lines: case study of a large database considering spatial componentsPublication . Freiria, Susana; Sousa, NunoThis article presents a large case study of fare evasion on bus lines in the city of Lisbon, Portugal, a common problem in dense urban areas. Focus is put on geographic factors, and an analysis is carried out using a generalized spatial two-step least-squares regression (GS2SLS). The large database, spanning one year of fare evasion reports, made it possible to segregate the analysis according to type of day (workday or weekend) and time period (rush hours, nighttime, etc.). Results show that indeed the type of day and time period lead to considerable differences between the seven models analyzed. It was found that the number of inspection actions is the strongest predictor of evasion, with geographic factors also playing a role in predicting fare evasion. Consideration of this spatial component made it possible to find moderate evidence for dissuasive effects of inspection actions in some models and of pockets of evasive tendencies in other models, which appear in the statistical error term. Interestingly, and contrary to other studies, age was found to have almost no influence on the propensity to evade fares. From a managerial point of view, this study highlights the importance of running inspection actions systematically and closely monitoring their outcomes. From a more theoretical standpoint, it suggests further research on geographic factors is needed to fully understand spatial patterns of evasion.