TEDE Coleção:
http://www.bdtd.uerj.br/handle/1/3665
2024-04-20T10:05:09ZIntegração de tecnologias embarcadas controladas por Inteligência Artificial: uma aplicação para apoiar o tratamento de fobias
http://www.bdtd.uerj.br/handle/1/21471
Título: Integração de tecnologias embarcadas controladas por Inteligência Artificial: uma aplicação para apoiar o tratamento de fobias
Autor: Jambo, Claudio Henrique Medaber
Primeiro orientador: Costa, Rosa Maria Esteves Moreira da
Abstract: In recent years, the health area has received technological contributions that provide
support for diagnostic practices, monitoring and treatment of different disorders and diseases,
mainly combining various techniques of Artificial Intelligence, Virtual Reality and
Mobile Computing. There are many challenges to integrate these technologies and provide
solutions that consider the automation of processes, the simplification of interaction
between professionals and patients, the low price of equipment, the individualization of
use, mobility and the use of Artificial Intelligence strategies. Aiming to overcome limitations
of two previous works, which applied technological combinations in the desensitization
of stress and phobias, this work aims to develop a technological combination that integrates
an autonomous and low-cost virtual environment embedded in an ESP32 board,
with multi-agent control, with support for natural language communication, to be used
in the Treatment by Exposure in Virtual Environments - VRET in the area of Clinical
Psychology, more specifically related to Anxiety Disorders. Low-cost virtual reality glasses
were used, with visualization on a smartphone. The prototype, called PhobIA 3DS,
is controlled by multi-agents that have modules for capturing physiological signals (heart
rate); uses natural language to obtain the level of anxiety perceived by the patient;
considers these two pieces of information in a Fuzzy module, which, in turn, generates a
response on the calculated level of anxiety; and controls and changes the display of specific
scenarios for each level of anxiety. Finally, the system was evaluated by a group of
experienced professionals, to verify aspects of the interface, relevance and usability. The
data obtained by the evaluation showed positive results and good prospects for using the
system in real activities
Instituição: Universidade do Estado do Rio de Janeiro
Tipo do documento: Dissertação2023-07-31T00:00:00ZSobre coloração total dos grafos circulantes
http://www.bdtd.uerj.br/handle/1/21045
Título: Sobre coloração total dos grafos circulantes
Autor: Alves Junior, Mauro Nigro
Primeiro orientador: Nobrega, Diana Sasaki
Abstract: A circulant graph Cn(d1, d2, • • • , dl) with 1 ≤ di ≤ bn 2 c, where di 6= dj , has vertex set V = {v0, v1, • • • , vn−1} and edge set E = Sl i=1 Ei , where Ei = {e i 0 , ei 1 , • • • , ei n−1} and e i j = vjvj+di , where the indexes of the vertices are considered modulo n. An edge of Ei is called edge of length di . A k-total coloring of a graph G is an assignment of k colors to the vertices and edges (elements) of G so that adjacent or incident elements have different colors. The total chromatic number of G is the smallest integer k for which G has a k-total coloring. The well known Total Coloring Conjecture states that the total chromatic number of a graph is either ∆(G) + 1 or ∆(G) + 2, where ∆(G) is the maximum degree of G. Graphs with χ 00(G) = ∆(G) + 1 are known as Type 1, and graphs with χ 00(G) = ∆(G) + 2 are known as Type 2. Some classical circulant graphs, such as the cycle graphs Cn ' Cn(1), the complete graphs Kn ' Cn(1, 2, ..., bn/2c), and the complete bipartite graphs Kn,n ' C2n(1, 3, 5, ..., k), where k is the biggest odd number such that k ≤ n, have their total chromatic number determined. Furthermore, the total chromatic number of every cubic circulant graph C2n(d, n) was determined by Hackmann and Kemnitz in 2004. There are many results in the well known powers of cycles graphs, an infinite family of circulant graphs Cn(1, 2, ..., k). In 2003, Campos and de Mello proved that Cn(1, 2) is Type 1, except for graph C7(1, 2) which is Type 2, and they conjectured that Cn(1, 2, ..., k), with 2 ≤ k ≤ bn/2c, is Type 2 if and only if n is odd and k < n/3 − 1. Recently, it was proved that this conjecture holds for k = 3 and k = 4. In 2008, Khennoufa and Togni proved that every 4-regular circulant graph C5p(1, k) is Type 1, for any positive integer p and k < 5p/2 with k ≡ 2 mod 5 or k ≡ 3 mod 5; and proved that C6p(1, k) is Type 1, for p ≥ 3 and k < 3p with k ≡ 1 mod 3 or k ≡ 2 mod 3. Furthermore, they vericated some particular cases with the help of the computer. In the same paper, Khennoufa and Togni conjectured that except for a finite number of Type 2, 4-regular circulant graphs are all Type 1. In this work, we studied all the results that envolved the state of art about total coloring of circulant graphs. Futhermore, we contribute to this conjecture by determining the total chromatic number of all graphs of the following three infinite families of 4-regular circulant graphs: Cn(2k, 3), k ≥ 1 and n = (8µ + 6λ)k, for non negative integers µ and λ; C3n(1, 3), for n > 1; and C3λp(1, p), λ ≥ 1 and p ≡ 0 mod 3, suggesting that the conjecture has a positive answer.
Instituição: Universidade do Estado do Rio de Janeiro
Tipo do documento: Dissertação2021-02-19T00:00:00ZInA2rMS: Instrumento de apoio à Avaliação da Arquitetura de MicrosServiço
http://www.bdtd.uerj.br/handle/1/20804
Título: InA2rMS: Instrumento de apoio à Avaliação da Arquitetura de MicrosServiço
Autor: Joppert, Angelo
Primeiro orientador: Oliveira, Marcelo Schots de
Abstract: The demand for more efficient and systematic approaches to system development is
constantly increasing. Some architectural styles have emerged as responses to traditional
monolithic approaches to system construction. As a result, large companies have been
transitioning from Monolithic Architecture (MOA) to Microservices Architecture (MSA).
MSA has gained prominence in the last decade due to its suitability for cloud-native
technologies and its distributed nature. The MSA can provide several benefits for organizations,
such as improvement in feature management and reuse, agility, maintenance,
deployment, and delivery processes, enhancing cost efficiency related to the software lifecycle.
Due to its complexity, MSA becomes challenging for architects and software
developers. Considering its distributed nature, its development requires a clear understanding
of the necessary features and the resources involved to implement them. In light
of this scenario, the conceptual analysis of the features in relation to the domain, weighing
their relevance and identifying their relationships, can support the evaluation of the compliance
to this architecture’s features. In this sense, this work aims to provide a tool to
support the analysis and evaluation of microservices (MS) concerning their adherence to
the features of MSA, allowing the degree of compliance with these features to be expressed
through an information visualization strategy. To achieve this objective, an approach is
proposed that consists of a features model, a glossary of terms, an evaluation questionnaire
integrated with a data visualization technique, to support architects and developers
during the evaluation of MS. Studies were conducted with the purpose of verifying the
suitability and applicability of the approach. The results provide positive evidence regarding
the adoption of the approach in the processes of building and maintaining MS.
It is believed that this work can holistically assist architects and development teams in
understanding and evaluating the diverse features involved in MSA, enhancing the overall
development process
Instituição: Universidade do Estado do Rio de Janeiro
Tipo do documento: Dissertação2023-08-02T00:00:00ZPerceptron Multicamadas com Backpropagation para Análise de Crédito
http://www.bdtd.uerj.br/handle/1/20780
Título: Perceptron Multicamadas com Backpropagation para Análise de Crédito
Autor: Figueiredo, Raphael Silva de
Primeiro orientador: Wedemann, Roseli Suzi
Abstract: The use of tools that use artificial intelligence is already quite present in our lives. Techniques such as the artificial neural network and other forms of machine learning have been growing year after year, helping us in many areas of knowledge, such as medicine, astronomy, economics, engineering, among others. In this work, we will study multilayer perceptron neural networks with the backpropagation learning algorithm, showing step by step how it was developed, its implementation in C language and applying it to a prac- tical problem of classifying debtors of the BNDES (Banco Nacional do Desenvolvimento Econ^omico e Social).
Instituição: Universidade do Estado do Rio de Janeiro
Tipo do documento: Dissertação2019-02-13T00:00:00Z