DigitUMa
University of Madeira Institutional Repository
Recent Submissions
Definição da estratégia de participação em feiras: parceria e/ou cooperação entre entidades públicas e privadas
Publication . Pereira, Bárbara Isabel Moniz; Teles, Susana; Almeida, António Manuel Martins de
O turismo desempenha um papel fundamental no desenvolvimento económico e
social de Portugal, assumindo particular relevância na Região Autónoma da Madeira
(RAM), onde contribui de forma estratégica para a criação de emprego, a valorização dos
recursos patrimoniais e o dinamismo económico local. Assim, este estudo tem como
objetivo principal analisar o efeito da participação em feiras de turismo no desempenho
das empresas hoteleiras da RAM, dando especial ênfase ao papel das parcerias público privadas (PPP) como apoio estratégico.
A investigação seguiu uma abordagem quantitativa, aplicando um inquérito junto
de representantes de empresas hoteleiras da região, com o propósito de compreender as
motivações para participar em feiras, os seus benefícios e desafios, e ainda o nível de
satisfação com esses eventos. Além disso, foi analisado o contributo das PPP para reforçar
a visibilidade, a competitividade e a capacidade de internacionalização das empresas
participantes.
Os resultados evidenciam a relevância das PPP na promoção de estratégias
colaborativas eficazes, impactando positivamente a presença das empresas em feiras
nacionais e internacionais. Este trabalho contribui para o avanço do conhecimento
científico ao propor e testar um modelo conceptual centrado na cooperação institucional
no setor turístico, oferecendo igualmente recomendações práticas dirigidas a gestores
públicos, empresários e decisores políticos.
Emotional regulation assessment via multi-biosignal processing in a VR environment for neurorehabilitation
Publication . Lima, Rodrigo Olival; Bermúdez i Badia, Sergi; Gamboa, Hugo Filipe Silveira; Cameirão, MÛnica da Silva
Emerging immersive technologies and physiological computing capabilities are opening
promising pathways for emotion recognition and regulation, with growing relevance in
fields such as affective computing, neurorehabilitation, and human-computer interaction.
Through four exploratory studies, this thesis investigates how virtual reality, biofeed back, and machine learning can be combined to recognize and regulate users’ emotional
states in real time.
First, a machine learning pipeline was developed to classify emotional states using
physiological signals collected in immersive and non-immersive virtual reality conditions.
Results showed that immersion had a limited impact on subjective emotional ratings, while
user-dependent models significantly outperformed user-independent ones, highlighting
the importance of personalization in emotion recognition.
The second study validated this pipeline in individuals with Alzheimer’s, revealing
that emotional reactivity is partially preserved across severity levels. Classification models
successfully distinguished between emotional states, healthy and Alzheimer’s participants,
and even Alzheimer’s severity levels, underscoring the pipeline’s clinical relevance and
generalization.
The third study introduced a nature-based virtual reality environment, the Virtual Lev ada, which used real-time adaptation to users’ physiological stress levels via a biofeedback
mechanism. This study also implemented and evaluated real-time retraining strategies
for the stress classification model, addressing temporal drift and improving model robust ness. Although biofeedback effects were not statistically significant, both adaptive and
non-adaptive groups reported reduced physiological arousal and anxiety, supporting the
environment’s calming and restorative potential.
Finally, the fourth study improved the adaptive virtual reality system by integrating
online stress predictions and online model retraining. Results demonstrated improved
prediction stability over time and significant reductions in state anxiety, particularly in
individuals with elevated stress levels.
In conclusion, these findings validate the feasibility and effectiveness of progressively
adaptive, personalized virtual reality systems for emotion recognition and regulation.
This work contributes with novel insights into how online physiological monitoring and
ML adaptation can enhance emotional self-regulation, offering promising directions for
affective technologies development and mental health interventions.
Massively parallel GPU acceleration of population-based optimization metaheuristics: application to the solution large-scale systems of nonlinear equations
Publication . Silva, Bruno Miguel Pereira da; Lopes, Luiz Carlos Guerreiro; Mendonça, Fábio Rúben Silva
High-dimensional problems, such as large-scale Systems of Nonlinear Equations, are
challenging due to their complexity and nonlinear solution spaces. Population-based
optimization metaheuristics, such as Particle Swarm Optimization and Gray Wolf
Optimizer, can offer effective approaches. However, their computational demands
often exceed the capacity of traditional methods, particularly when addressing these
problems at large scales.
To address these challenges, parallelization constitutes a promising strategy. Due
to the massive parallel processing capabilities, a Graphics Processing Unit (GPU)
is well-adapted to the acceleration of population-based metaheuristic optimization
algorithms. Thus, employing GPU parallelism can substantially reduce computational
time and enable the solution of larger and more complex problems that would be
impractical on conventional Central Processing Units (CPUs).
GPU-based parallelization of metaheuristic optimization algorithms faces several
challenges due to algorithmic diversity and heterogeneous hardware architectures.
Different metaheuristics exhibit distinct computational patterns, memory access
requirements, and degrees of inherent parallelism, which complicates efficient mapping
to GPU architectures. Moreover, variations in GPU hardware can substantially affect
performance, often requiring algorithm-specific adaptations and hardware-aware
optimizations to fully exploit GPU resources.
This research proposes GPU-based parallelization strategies for population-based
metaheuristic algorithms to enhance performance on large-scale, high-dimensional
optimization problems. It uses GPU parallelism to manage increasing problem sizes
while preserving convergence behavior and solution quality. A central goal is a
hardware-agnostic model that enables scalable acceleration across diverse computa tional environments, providing a general framework for GPU-based metaheuristic
acceleration applicable to various algorithmic paradigms and problem domains.
Experimental results indicate that GPU-accelerated metaheuristics using the
proposed framework substantially outperform their sequential counterparts, achieving
significant speedups. The framework scaled effectively across ten population-based
algorithms and ten benchmark problems of increasing dimensionality, utilizing five
GPU models, including both consumer-grade and professional-grade hardware. In
multi-GPU tests, the framework exhibited superlinear speedup in certain cases.
This study highlights the value of modular, reproducible frameworks for GPU based metaheuristics and provides a base for future research in high-dimensional,
computationally intensive optimization.
User profiling with feature selection and explainability: essays on three case studies across different domains
Publication . Freitas, Diogo Nuno Teixeira; Teixeira Freitas, Diogo Nuno; Dias, Fernando Manuel Rosmaninho Morgado Ferrão; Fermé, Eduardo Leopoldo
User profiling is the process of constructing a structured representation of the user
within a system. This representation includes information such as preferences, behaviors,
and characteristics. Based on the profile, the system can recommend services and products
or, in this work, suggest actions. Machine learning methods are commonly used to this
end, as they can identify complex patterns among large numbers of attributes.
However, not all attributes are relevant. High-dimensional datasets often contain
irrelevant, redundant, or noisy features that obscure valuable patterns and reduce model
accuracy. To address this, dimensionality reduction techniques—particularly feature
selection—are essential. Equally important is the ability to explain a model’s output, since
understanding why a model produces a given outcome builds trust and clarifies which
steps can change an undesirable situation.
This thesis applies feature selection, explainability, causal discovery, and machine
teaching techniques to user profiling. The goal is to support decision-making by identi fying the most relevant features, clarifying causal mechanisms, and ensuring that stake holders understand why recommendations are made. Specifically, we investigate the
mRMR (minimum-Redundancy-Maximum-Relevance) method for feature selection, ex amine explainability strategies such as feature importance analysis and counterfactuals,
apply causal discovery to map cause-and-effect relationships, and use machine teaching
to explore profile simplification.
We apply this approach in four domains: (i) Marine litter: developing static profiles to
identify those who could benefit from literacy interventions; (ii) Football injuries: building
predictive models based on player profile dynamics to forecast risk; (iii) Energy poverty:
designing models, using counterfactuals, and applying causal discovery to understand
health–poverty links; and (iv) Concept complexity: using machine teaching to study
profile simplification.
These applications show how profiling can deliver targeted literacy interventions,
prevent sports injuries, inform preventive policies in energy poverty, and improve the
efficiency of user representations and concept learnability.
Fatores de atração de nómadas digitais: o caso da Região Autónoma da Madeira
Publication . Fernandes, Filipa Teixeira; Martins, António Miguel Valente; Cró, Susana Raquel Granito
O nomadismo digital tem vindo a consolidar-se como um segmento turístico emergente, que
combina simultaneamente lazer e trabalho remoto. A Região Autónoma da Madeira (RAM) tem
procurado afirmar-se neste contexto ao implementar políticas e iniciativas específicas para
atrair este segmento. No entanto, os fatores que influenciam a duração da estadia deste
segmento permanecem pouco explorados, justificando a pertinência da presente investigação.
O presente estudo tem como objetivo analisar os fatores que influenciam a duração da
estadia dos nómadas digitais na RAM, contribuindo para a literatura sobre este segmento
turístico e estilo de vida emergente. Para tal, adotou-se uma abordagem quantitativa, baseada
na aplicação de um questionário aos nómadas digitais presentes na região durante o período em
estudo. Os dados recolhidos foram analisados através de três modelos estatísticos
complementares conferindo robustez à análise.
Os resultados encontrados revelam que, ao contrário do turista tradicional, o
comportamento do nómada digital não é impulsionado por variáveis sociodemográficas ou por
fatores culturais. Em contrapartida, a duração da estadia mostrou-se positivamente associada a
fatores práticos e estruturais, como custo de vida, qualidade da internet e hospitalidade. Estes
resultados evidenciam que este segmento perceciona o destino como uma base funcional para
viver e trabalhar em vez de apenas um espaço de visita temporária.
O estudo contribui para a literatura académica ao testar variáveis e metodologias
tipicamente aplicadas no estudo do segmento turístico convencional a um contexto emergente,
preenchendo uma lacuna na literatura. Os resultados obtidos fornecem uma base para o
desenvolvimento de estratégias para a gestão do destino, sublinhando a importância do
investimento em infraestruturas e serviços que garantam a atratividade, retenção e
sustentabilidade deste segmento, mas também a criação de um ambiente que promova o
sentimento de pertença de forma a maximizar o potencial económico e social deste segmento.