CISeD - CENTRO DE ESTUDOS EM SERVIÇOS DIGITAIS
URI permanente desta comunidade:
Navegar
Percorrer CISeD - CENTRO DE ESTUDOS EM SERVIÇOS DIGITAIS por título
A mostrar 1 - 10 de 140
Resultados por página
Opções de ordenação
- Adaptive and Scalable Database Management with Machine Learning Integration: A PostgreSQL Case StudyPublication . Abbasi, Maryam; Bernardo, Marco V.; Vaz, Paulo; Silva, José; Martins, Pedro; ANTUNES VAZ, PAULO JOAQUIM; Silva, JoséThe increasing complexity of managing modern database systems, particularly in terms of optimizing query performance for large datasets, presents significant challenges that traditional methods often fail to address. This paper proposes a comprehensive framework for integrating advanced machine learning (ML) models within the architecture of a database management system (DBMS), with a specific focus on PostgreSQL. Our approach leverages a combination of supervised and unsupervised learning techniques to predict query execution times, optimize performance, and dynamically manage workloads. Unlike existing solutions that address specific optimization tasks in isolation, our framework provides a unified platform that supports real-time model inference and automatic database configuration adjustments based on workload patterns. A key contribution of our work is the integration of ML capabilities directly into the DBMS engine, enabling seamless interaction between the ML models and the query optimization process. This integration allows for the automatic retraining of models and dynamic workload management, resulting in substantial improvements in both query response times and overall system throughput. Our evaluations using the Transaction Processing Performance Council Decision Support (TPC-DS) benchmark dataset at scale factors of 100 GB, 1 TB, and 10 TB demonstrate a reduction of up to 42% in query execution times and a 74% improvement in throughput compared with traditional approaches. Additionally, we address challenges such as potential conflicts in tuning recommendations and the performance overhead associated with ML integration, providing insights for future research directions. This study is motivated by the need for autonomous tuning mechanisms to manage large-scale, hetero geneous workloads while answering key research questions, such as the following: (1) How can machine learning models be integrated into a DBMS to improve query optimization and workload management? (2) What performance improvements can be achieved through dynamic configuration tuning based on real-time workload patterns? Our results suggest that the proposed framework significantly reduces the need for manual database administration while effectively adapting to evolving workloads, offering a robust solution for modern large-scale data environments.
- Agile-based Requirements Engineering for Machine Learning: A Case Study on Personalized NutritionPublication . Cunha, Carlos; Oliveira, Rafael; Duarte, RuiRequirements engineering is crucial in developing machine learning systems, as it establishes the foundation for successful project execution. Nevertheless, incorporating requirements engineering approaches from traditional software engineering into machine learning projects presents new challenges. These challenges arise from replacing the software logic derived from static software specifications with dynamic software logic derived from data. This paper presents a case study exploring an agile requirement engineering approach popular in traditional software projects to specify requirements in machine learning software. These requirements allow reasoning about the correctness of software and design tests for validation. The absence of software specification in machine learning software is offset by employing data quality metrics, which are assessed using cutting-edge methods for model interpretability. A case study on personalized nutrition and physical activity demonstrated the adequacy of user stories and acceptance criteria format, popular in agile projects, for specifying requirements in the machine learning domain.
- AI-Powered Data Management to Optimize Data Collection and Processing in a Painting LaboratoryPublication . Pereira, Maria Teresa Ribeiro; Pereira, Marisa João Guerra; Tavares, Miguel Guedes; Guimarães, André; Vilarinho, HermilioIndustrial laboratories often remain under-digitized compared to production lines, creating a gap between data acquisition and analytical intelligence, critical for advanced quality control. This study addresses this gap by proposing and validating a novel framework that combines Low-Code digitalisation tools with Machine Learning (ML) and Causal Inference to optimise data collection and analysis in an automotive painting laboratory. A Microsoft Power Apps-based platform was developed in order to digitalise all measurement records, eliminating manual transcription errors (previously ≈ 40.01%) and reducing data-handling time by up to 34% of an operator’s shift, while enabling centralised, traceable storage and Power BI integration. Four datasets were used to assess predictive capacity with Random Forest, XGBoost and Neural Networks; Random Forest consistently provided the most stable results—Mean Absolute Error (MAE) of 0.972, Mean Absolute Percentage Error (MAPE) of 16.45%, and Root Mean Square Error (RMSE) of 1.307. Causal models (Linear Regression, DoWhy, Causal Forest, Double Machine Learning) consistently identified ultrafiltrate I solid content of the electrodeposition process as a dominant causal factor for defects. This study provides a novel framework that bridges digitalisation and ML-based causal reasoning in laboratory settings, offering a scalable approach that can be extended and replicated in other industrial sectors, aiming to develop smart, data-driven quality control systems.
- AI-Powered Data Management to Optimize Data Collection and Processing in a Painting LaboratoryPublication . Pereira, Teresa; Oliveira, Marisa; Tavares, Miguel; Guimarães, AndréIndustrial laboratories often remain under-digitized compared to production lines, creating a gap between data acquisition and analytical intelligence, critical for advanced quality control. This study addresses this gap by proposing and validating a novel framework that combines Low-Code digitalisation tools with Machine Learning (ML) and Causal Inference to optimise data collection and analysis in an automotive painting laboratory. A Microsoft Power Apps-based platform was developed in order to digitalise all measurement records, eliminating manual transcription errors (previously ≈ 40.01%) and reducing data-handling time by up to 34% of an operator’s shift, while enabling centralised, traceable storage and Power BI integration. Four datasets were used to assess predictive capacity with Random Forest, XGBoost and Neural Networks; Random Forest consistently provided the most stable results—Mean Absolute Error (MAE) of 0.972, Mean Absolute Percentage Error (MAPE) of 16.45%, and Root Mean Square Error (RMSE) of 1.307. Causal models (Linear Regression, DoWhy, Causal Forest, Double Machine Learning) consistently identified ultrafiltrate I solid content of the electrodeposition process as a dominant causal factor for defects. This study provides a novel framework that bridges digitalisation and ML-based causal reasoning in laboratory settings, offering a scalable approach that can be extended and replicated in other industrial sectors, aiming to develop smart, data-driven quality control systems.
- Airbnb research: an analysis in tourism and hospitality journalsPublication . Andreu, Luisa; Bigne, Enrique; Amaro, Suzanne; Palomo, JesúsPurpose – The purpose of this study is to examine Airbnb research using bibliometric methods. Using research performance analysis, this study highlights and provides an updated overview of Airbnb research by revealing patterns in journals, papers and most influential authors and countries. Furthermore, it graphically illustrates how research themes have evolved by mapping a co-word analysis and points out potential trends for future research. Design/methodology/approach – The methodological design for this study involves three phases: the document source selection, the definition of the variables to be analyzed and the bibliometric analysis. A statistical multivariate analysis of all the documents’ characteristics was performed with R software. Furthermore, natural language processing techniques were used to analyze all the abstracts and keywords specified in the 129 selected documents. Findings – Results show the genesis and evolution of publications on Airbnb research, scatter of journals and journals’ characteristics, author and productivity characteristics, geographical distribution of the research and content analysis using keywords. Research limitations/implications – Despite Airbnb having a history of 10 years, research publications only started in 2015. Therefore, the bibliometric study includes papers from 2015 to 2019. One of the main limitations is that papers were selected in October of 2019, before the year was over. However, the latest academic publications (in press and earlycite) were included in the analysis. Originality/value – This study analyzed bibliometric set of laws (Price’s, Lotka’s and Bradford’s) to better understand the patterns of the most relevant scientific production regarding Airbnb in tourism and hospitality journals. Using natural language processing techniques, this study analyzes all the abstracts and keywords specified in the selected documents. Results show the evolution of research topics in four periods: 2015-2016, 2017, 2018 and 2019
- An automated closed-loop framework to enforce security policies from anomaly detectionPublication . Henriques, João; Caldeira, Filipe; Cruz, Tiago; Simões, PauloDue to the growing complexity and scale of IT systems, there is an increasing need to automate and streamline routine maintenance and security management procedures, to reduce costs and improve productivity. In the case of security incidents, the implementation and application of response actions require significant efforts from operators and developers in translating policies to code. Even if Machine Learning (ML) models are used to find anomalies, they need to be regularly trained/updated to avoid becoming outdated. In an evolving environment, a ML model with outdated training might put at risk the organization it was supposed to defend. To overcome those issues, in this paper we propose an automated closed-loop process with three stages. The first stage focuses on obtaining the Decision Trees (DT) that classify anomalies. In the second stage, DTs are translated into security Policies as Code based on languages recognized by the Policy Engine (PE). In the last stage, the translated security policies feed the Policy Engines that enforce them by converting them into specific instruction sets. We also demonstrate the feasibility of the proposed framework, by presenting an example that encompasses the three stages of the closed-loop process. The proposed framework may integrate a broad spectrum of domains and use cases, being able for instance to support the decide and the act stages of the ETSI Zero-touch Network & Service Management (ZSM) framework.
- An Importance-Performance Matrix Analysis of the Factors Influencing US Tourists to Use AirbnbPublication . Palmer, Alicia; Amaro, Suzanne; Andreu, LuisaAfter 10 years of Airbnb and its significant growth, this research is conducted to identify which factors really matter to Airbnb users today. It provides an integrated approach, since it aggregates factors that other studies have considered separately. With a sample of 101 US respondents, the partial least squares structural equation modelling results evidence that the only factors that influence future intentions to use Airbnb are its unique and varied accommodations and satisfaction. Enjoyment was also found to have a positive effect on Satisfaction. The study further investigates these factors with an Importance Performance map, providing guidance for the prioritization of managerial activities of high importance for futures intentions to use Airbnb.
- An integrated and interoperable AutomationML-based platform for the robotic process of metal additive manufacturingPublication . Babcinschi, Mihail; Freire, Bernardo; Ferreira, Lucía; Señaris, Baltasar; Vidal, Felix; Vaz, Paulo; Neto, PedroIncreasingly, industry is looking to better integrate their industrial processes and related data. Interoperability is key since the organizations need to share data between them, between departments and the different stages of a given technological process. The problem is that many times there are no standard data formats for data exchange between heterogeneous engineering tools. In this paper we present an integrated and interoperable AutomationML-based platform for the robotic process of metal additive manufacturing (MAM). Data such as the MAM robot targets and process parameters are shared and edited along the different sub-stages of the process, from Computer-Aided Design (CAD), to path planning, to multiphysics simulation, to robot simulation and production. The AutomationML neutral data format allows the implementation of optimization loops connecting different sub-stages, for example the multi-physics simulation and the path planning. A practical use case using the Direct Energy Deposition (DED) process is presented and discussed. Results demonstrated the effectiveness of the proposed AutomationML-based solution.
- Analysis of the factors that influence the success of relationship marketing in academic libraries.Publication . Figueiredo, Elisabeth; Pereira, Paulo; Ribeiro, Célia; Passos, Clotilde; Antunes, JoaquimPurpose: This study aims to identify the main dimensions of relationship marketing that have the greatest impact on users’ satisfaction and loyalty in academic libraries. Design/methodology/approach: A quantitative study was conducted. Data were collected through a questionnaire survey, directed to the users of four libraries of Universidade Católica Portuguesa, which resulted a total of 292 valid responses. The necessary sample size for the completion of this study was calculated based on Daniel's calculator Soper. Statistical Package for Social Sciences (SPSS 27.0) software and structural equation methods were use to test the proposed model. Originality: This study contributes to the existing literature by examining the specific dimensions of relationship marketing that are most influential in the context of university libraries. While prior research has explored relationship marketing in various contexts, this study focuses on the unique environment of academic libraries, thereby providing novel insights into enhancing user satisfaction and fostering loyalty among library users. Findings: The results show that the relationship marketing dimensions that contribute the most to users’ satisfaction and loyalty are: "relational behaviour", "understanding needs" and "quality of services". Additionally, the study identifies the positive relationship between user satisfaction and loyalty, emphasizing the critical role of relationship marketing in fostering long-term user loyalty in the academic library setting. Theoretical/methodological contributions: Increase of scientific knowledge through the validation of a relationship marketing model, through which its implications for the management of university libraries are evidenced. Practical implications: Understanding the main dimensions of relationship marketing that impact user satisfaction and loyalty allows library managers to adapt their strategies and services to motivate and attract users to their spaces and to better meet their needs.
- Aplicação da metodologia DMAIC em uma empresa produtora de componentes de borrachaPublication . Almeida, Ricardo; ANTUNES VAZ, PAULO JOAQUIM; Gomes da Silva, Rosa Maria; AlmeidaIntrodução: Nas últimas décadas, foram feitos grandes desenvolvimentos na indústria. As empresas aprimoraram-se para fazer o produto final com mais qualidade e com grande redução de custos. As metodologias Lean foram implementadas em todos os tipos de indústrias e negócios, rompendo com o tipo de produção que se praticava na época, que se baseava em grandes volumes e pouco flexíveis. As metodologias Lean começaram quando os funcionários e engenheiros da Toyota começaram a desenvolver procedimentos e ferramentas para permitir a produção lean, com desperdício zero e sistemas de produção altamente flexíveis. Objetivo:A reorganização do layout do departamento de manutenção, assim como, a melhoria do processo de gestão das spare partse a criação de fluxos para a reparação de equipamentos e ferramentas. Métodos: A ferramenta utilizada foi o DMAIC, esta subdivide o processo de resolução de problemas em cinco etapas, tais como: Definir, Medir, Analisar, Melhorar, Controlar. Resultados: Com a aplicação desta ferramenta foi possível uma redução do número de deslocações e da distância percorrida, (que por sua vez, permitiu também a diminuição do tempo necessário para a sua realização) deste modo o tempo necessário para a sua realização também diminuiu. As spare parts estão mais organizadas, cada bancada de trabalho possui as peças de substituição de maior consumo. A pontuação obtida nas auditorias 5’S também apresentaram um aumento face aos resultados obtidos antes da intervenção. Conclusão: Conclui-se que a causa raiz e as soluções definidas impactaram positivamente a eliminação da causa e problema iniciais.