PPGI https://repositorio.ufpb.br/jspui/handle/tede/5439 2026-09-19T08:06:57Z 2026-09-19T08:06:57Z Silva, Marcell Bruno Sousa e https://repositorio.ufpb.br/jspui/handle/123456789/38900 2026-09-05T06:13:02Z 2025-12-15T00:00:00Z

Título: Contribuições ao estudo de sistemas de detecção de intrusão em hosts para ataques de dia zero Autor(es): Silva, Marcell Bruno Sousa e Orientador: Fonseca, Iguatemi Eduardo da Abstract: In the context of the ubiquitous use of digital media and data in contemporary society, cybersecurity emerges as one of the most critical areas. In this context, computer intrusion detection systems emerge as an essential tool for ensuring the security of digital data and media. This work aimed to evaluate the performance of LSTM, Transformer, and Kohonen Map (SOM) techniques in identifying zero-day attacks on hosts. The evaluated models were employed within the unsupervised classification paradigm, which does not use labels for classes in the training process. The unsupervised paradigm is capable of identifying new attacks and minimizing the reduced availability of attack data in the face of the immense amount of data originating from normal host operations. The performance measures adopted were intrusion detection rate, false alarm rate, number of parameters, and number of floating-point operations of the adopted models, in order to evaluate not only the final effectiveness of each model but also to consider its computational complexities. It was demonstrated that the technique that presented the best performance was the Transformer with an AUC (Area Under the Curve) of 0.990, however, the SOM presented lower computational complexity with an intermediate AUC. Editor: Universidade Federal da Paraíba Tipo: Dissertação

2025-12-15T00:00:00Z Ramos, Erikson Carlos https://repositorio.ufpb.br/jspui/handle/123456789/38886 2026-09-04T06:12:28Z 2025-08-29T00:00:00Z

Título: Detecção automática de Aanomalias baseada em Difusão Geométrica Markoviana Autor(es): Ramos, Erikson Carlos Orientador: Souza, Leandro Carlos de Abstract: Automatic anomaly detection is essential in domains such as healthcare, finance, and cybersecurity, where small deviations may indicate fraud, failure, or imminent risk. This study introduces an unsupervised method called Anomaly Detection based on Marko- vian Geometric Diffusion (DA-DGM), applicable to one-dimensional and multi-dimensional datasets. The technique employs geometric diffusion to identify nonlinear relationships among instances; in multi-dimensional data, the scale parameter is tuned via Shannon entropy using a global search followed by local refinement, ensuring adaptability to the data context. The main contribution lies in integrating the sensitivity of Markovian geometric diffusion with automatic scale-parameter tuning based on Shannon entropy, thereby eliminating reliance on manual hyperparameter adjustment and providing a di- rectly interpretable anomaly metric grounded in self-transition probability. In evaluations on synthetic and real datasets, DA-DGM matches or surpasses classical methods such as KNN (k-Nearest Neighbors) and LOF (Local Outlier Factor), and achieves competitive performance against IForest (Isolation Forest) in terms of area under the ROC curve (AUC-ROC), with particular advantages in scenarios of complex geometry and low ano- maly prevalence. These results highlight DA-DGM’s potential for critical applications and reinforce the usefulness of diffusion-based approaches in data mining. Editor: Universidade Federal da Paraíba Tipo: Dissertação

2025-08-29T00:00:00Z Morais, Rafael Sobral de https://repositorio.ufpb.br/jspui/handle/123456789/38492 2026-07-24T06:10:39Z 2026-01-30T00:00:00Z

Título: Exact and heuristic approaches for single machine scheduling with inventory constraints and sequence-dependent setup times Autor(es): Morais, Rafael Sobral de Orientador: Subramanian, Anand Abstract: This work addresses the single machine scheduling problem with release dates, sequence-dependent setup times, and inventory constraints, 1|rj , sij , inv|Cmax. Four mixed-integer programming formulations are proposed: position-indexed, arc- indexed, arc-inventory-indexed, and arc-time-indexed. Computational experiments were performed, and the results show that the arc-indexed formulation provides the best performance among the models. Building on this formulation, a branch-and-cut algorithm incorporating inventory capacity cuts is developed, along with a branch-cut-and-price algorithm that extends an existing branch-and-price framework for scheduling problems. Among the exact methods, the branch-cut-and-price approach solves the largest number of instances to optimality, while the branch-and-cut method attains the lowest average gaps. Two heuristic methods based on the Iterated Local Search framework are also pro- posed, using ruin-and-recreate and SISRs-inspired perturbation strategies. Both heuristics consistently generate feasible solutions, with the ruin-and-recreate variant achieving particularly competitive performance. Editor: Universidade Federal da Paraíba Tipo: Dissertação

2026-01-30T00:00:00Z Lemos, Diego Oliveira de https://repositorio.ufpb.br/jspui/handle/123456789/38252 2026-06-30T06:11:13Z 2026-02-12T00:00:00Z

Título: Detecção e análise de dark patterns em plataformas de apostas no Brasil Autor(es): Lemos, Diego Oliveira de Orientador: Siebra, Clauirton de Albuquerque Abstract: The rapid expansion of sports betting platforms in Brazil has brought about the prolifera- tion of manipulative design techniques, known as Dark Patterns, which can compromise user autonomy and decision-making. This dissertation proposes a hybrid methodology for the identification and analysis of these practices, integrating Natural Language Processing (NLP) techniques and structural interface inspection. The methodological path involved the investigation of 22 platforms operating in the national market during the period in which the research was conducted. For textual analysis, the BERTimbau model — a Transformer-based architecture fine-tuned for Brazilian Portuguese — was used, achieving a weighted average F1-Score of 0.82 in the classification of categories such as Urgency and Social Proof. Concurrently, the inspection of style sheets (CSS) revealed more than 8,000 occurrences of the opacity:0 property, signaling the presence of potentially hidden or obstructive interface elements. The results show that platforms adapt traditional e-commerce tactics to the betting context, exploring cognitive vulnerabilities through temporal pressure and programmed technical barriers. It is concluded that the sophisti- cation of these strategies, although detectable by automated methods, demands specific regulation regarding interface design in the sector, aiming for consumer protection and the preservation of mental health in the face of predatory business models. Editor: Universidade Federal da Paraíba Tipo: Dissertação

2026-02-12T00:00:00Z