http://repositorio.ufc.br/handle/riufc/22343 2026-05-30T10:35:55Z http://repositorio.ufc.br/handle/riufc/86301 Título: Análise de desempenho das tecnologias de carregamento de dados em bancos de dados pela linguagem Java Autor(es): Santos, Agatha Lindemberg Araújo dos Abstract: This work presents a comparative analysis of the main data persistence technologies used in Java applications: JDBC, Hibernate, JPA, and Spring Data JPA. Considering the growth in data volume and the need for increasingly efficient systems, it investigated how each technology influences performance, resource consumption, and development complexity. The research was guided by the following question: what are the advantages and disadvantages of each data manipulation approach in the context of modern Java applications? To answer this question, the study conducted a theoretical review of the different data access and persistence strategies, highlighting their technical characteristics, abstraction models, productivity, and impact on performance. Both low-level approaches, such as JDBC, and more abstract solutions, such as Hibernate, JPA, and Spring Data JPA, based on the ORM paradigm, which performs the mapping between Java language objects and relational database structures, were analyzed. The results indicate that JDBC offers greater control and raw performance, but requires more implementation and maintenance effort. ORM technologies, such as Hibernate and JPA, provide greater productivity and facilitate development, although they can introduce memory overflow and scalability problems in high-load scenarios. Spring Data JPA stands out for its simplicity and speed of development, but may suffer limitations when highly customized configurations or intensive operations are required. The analysis shows that the choice of the most appropriate technology depends directly on the application context, involving factors such as complexity of business logic, performance requirements, data volume, and the profile of the development team. It is concluded that no technology is universally superior, reinforcing the importance of well-founded architectural decisions. The study contributes to guiding professionals in selecting the most appropriate persistence approach and to fostering discussions about performance, scalability, and productivity in the Java ecosystem. Tipo: TCC 2026-01-01T00:00:00Z http://repositorio.ufc.br/handle/riufc/86300 Título: Átomos de confusão: uma revisão sistemática da literatura Autor(es): Furtado, Ana Carolina Freire Abstract: Code comprehension is a fundamental cognitive activity in software development, directly impacting system maintenance and quality. However, certain syntactic patterns, called “Atoms of Confusion” (AoCs), tend to lead developers to misinterpretations, even though they are technically correct. Given the dispersion of knowledge on the subject, this work presents a Systematic Literature Review (SLR) with the aim of consolidating the state of the art on ACs and their impacts on software engineering. The research was conducted based on a rigorous protocol (PICOC), resulting in the selection and analysis of 32 primary studies published between 2017 and 2025. The results indicate that the taxonomy of atoms has evolved from an initial focus on C/C++ to include managed (Java) and dynamic (JavaScript) languages, revealing that confusion is a universal phenomenon, but one shaped by the language paradigm. Empirical evidence confirms that the presence of ACs increases cognitive load, reading time, and the propensity for errors, although recent studies on mature Java systems suggest that cases in which an atom results in functional defects are rare. A technological gap was also identified in the absence of real-time detection tools (IDE plugins) and the emergence of Artificial Intelligence as a new frontier, acting both as a refactoring agent and as a potential propagator of confusion. It is concluded that effective mitigation of AoC requires a hybrid approach, combining education, active barriers in development pipelines, and the cautious use of AI-based coding assistants. Tipo: TCC 2026-01-01T00:00:00Z http://repositorio.ufc.br/handle/riufc/86299 Título: Análise comparativa de arquiteturas mobile para aplicações iOS: MVC, MVVM e VIP-C Autor(es): Souza, Anderson de Alencar Bezerra Abstract: Planning the development of mobile applications in the iOS ecosystem involves architectural decisions that can directly impact code organization, system evolution, and software quality over time. Among these decisions, the choice of software architecture plays a central role and must consider technical requirements, business objectives, and the intended growth of the project. In this context, the following research question arises: how do the MVC, MVVM, and VIP-C architectures influence software quality attributes, and in which project contexts is each architecture more suitable for iOS application development? The diversity of approaches adopted in practice reinforces the relevance of this investigation. The objective of this study is to conduct a comparative analysis of the Model–View–Controller (MVC), Model–View–ViewModel (MVVM), and View–Interactor–Presenter–Coordinator (VIP-C) architectures, identifying their impacts on quality attributes such as modularization, coupling, cohesion, data flow, testability, and maintainability, as well as to establish guidelines to support architectural decision-making across different project scenarios in the iOS context. To achieve this objective, a qualitative and analytical methodology was adopted, based on a review of scientific and technical literature, including empirical, conceptual, and applied studies. The selected works were analyzed comparatively, considering the distribution of responsibilities among components and the communication mechanisms between architectural layers. The results indicate that MVC is predominantly adopted in initial and low-complexity applications due to its structural simplicity and integration with native iOS frameworks; that MVVM is more suitable for small- to medium-sized projects, especially in the SwiftUI context, as it promotes a clearer separation between presentation logic and user interface; and that more segmented architectures, such as VIP-C, demonstrate superior performance in terms of modularization, testability, and maintainability in medium- and large-scale systems, at the cost of greater initial implementation effort. Tipo: TCC 2026-01-01T00:00:00Z http://repositorio.ufc.br/handle/riufc/85078 Título: Desenvolvimento de um Largue Language Model: uma estratégia de aprimoramento de transcrições de anamneses médicas Autor(es): Gonçalves, Yanna Torres Abstract: This work aims to explore the development of a Large Language Model (LLM) to enhance the transcription of patient records. The research involved two main stages: the first consisted of an initial benchmark of three language models (Phi3, Llama, and Mistral), evaluating their performance in the task of formatting anamneses, and the second focused on fine-tuning the Mistral model, based on qualitative and quantitative feedback from the evaluators. The Mistral model stood out in the benchmark due to its consistency and better results in agreement metrics, such as Kendall’s tau and Fleiss’ Kappa coefficients, although it was criticized for overloading details in its responses, impairing its objectivity. The Llama model performed the worst, with responses that were often irrelevant, while Phi3 showed moderate performance, but with issues in precision and excessive use of generic language. The fine-tuning of the Mistral model, despite attempts to correct the flaws, did not result in significant improvements, as the average evaluation score was slightly lower than the original model. The main contributions of this work include the comparison of language models for a specific task in the medical field and the analysis of the limitations of the fine-tuning process. The study also identified important limitations, such as data quality and potential bias among the evaluators, which may have influenced the results. As future work, the intention is to conduct longer training sessions, expand and diversify the dataset, and explore new machine learning approaches to further improve the model’s performance. Tipo: TCC 2024-01-01T00:00:00Z