Redes neurais para grafos e suas aplicações aos sistemas complexos

• Autor(a) principal: Carvalho, Guilherme Michel Lima de
• Data de Publicação: 2022
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da UFSCAR
• Texto Completo: https://repositorio.ufscar.br/handle/ufscar/16236
• topic: Redes neurais para grafos; Sistemas complexos; Aprendizado de máquina; Graph neural networks; Complex systems; Machine Learning; CIENCIAS EXATAS E DA TERRA::PROBABILIDADE E ESTATISTICA::ESTATISTICA::ANALISE DE DADOS; CIENCIAS EXATAS E DA TERRA::PROBABILIDADE E ESTATISTICA::ESTATISTICA::INFERENCIA EM PROCESSOS ESTOCASTICOS; CIENCIAS EXATAS E DA TERRA::MATEMATICA::MATEMATICA APLICADA

Complex systems are composed of several components that interact with each other. A natural approach for these types of systems is to use mathematical graph abstraction. In different contexts in the real world, it is possible to use complex network techniques to model these systems. In these systems, dynamic processes such as the spread of information and the spread of disease can occur. In this work we consider the use of artificial neural network techniques for graph-structured data in order to study the propagation of rumor in complex networks and the detection of community structures. For the proposed case of rumor, a model was developed based on graph neural networks for the porpuse of detected the source of the a rumour in graphs with community structure and for community dectection was evaluate the potential of graph neural networks in comparison to traditional methods of the network science.