Navegando por Assunto "Redes neurais artificiais"

Agora exibindo 1 - 3 de 3

Algoritmos bioinspirados – uma revisão sistemática da literatura no Brasil
(Centro Universitário do Estado do Pará, 2020) Gomes, Wesley Gabriel Pereira; Tork, Vinicius Martins; Soares, Lucas Leal; Souza, Alan Marcel Fernandes de
We present a systematic literature review on Brazil regarding bio-inspired algorithms. This is a study on the bibliographic review of articles published in federal universities in the last four years (2016-2020). This study searched the indexed literature in the bibliographies by the following Portuguese keywords in computer science: “bioinspiradas”, “metaheurísticas” and “inteligência artificial”. This undergraduate thesis presents a brief introduction in the subject of bio-inspired algorithms, meta-heuristics and artificial intelligence and subsequently the results of the published studies about bio-inspired algorithms were analyzed, evaluating which of them utilized the Ant-Colony Optimization (ACO).
Otimização de hiperparâmetros de redes neurais por inferência Bayesiana com enxame de partículas
(Centro Universitário do Estado do Pará, 2023-12-07) Barros, Lucas Lima de Aragão; Souza, Daniel Leal; Mollinetti, Marco Antônio Florenzano; http://lattes.cnpq.br/1246642827046945; http://lattes.cnpq.br/6059334260016388; Nascimento, Polyana Santos Fonseca; http://lattes.cnpq.br/6889523334917369; Oliveira, Roberto Célio Limão de; http://lattes.cnpq.br/4497607460894318
This work explores - through hyperparameter optimization techniques - the possibility of improving the effectiveness of neural networks, which have experienced widespread popularity in various fields. The focus is to evaluate a new approach to hyperparameter optimization, which is one of the main challenges in the development of these networks. Traditionally, the definition of these hyperparameters occurs stochastically or through mathematical calculations. The goal is to refine the process, allowing a more precise definition of values, with lower losses. These solutions are based on a review of the literature on hyperparameter optimization in Neural Networks (e.g., CNN, Fully Connected), addressing fundamental concepts and techniques such as Grid Search, Random Search, and SMBO. The methodology includes the choice of tools, cross-validation strategy, and specific search approaches, such as Gaussian Regression or grid search. The detailed experiments involve optimizing the hyperparameters of Neural Networks, presenting datasets, model configurations, training protocols, and quantitative results. Despite its limitations and the need for further studies, this work demonstrates the feasibility and potential of combining advanced techniques for hyperparameter optimization in Neural Networks, offering contributions and stimulating new research directions in the search for greater efficacy in various applications of these networks. The discussion analyzes the effects of hyperparameter settings, identifies limitations, and addresses possible reasons for the obtained results. Finally, the conclusion summarizes the main insights, specific contributions, and future directions, while the references ensure the work's foundation.
Previsão de demandas em uma rede de postos de combustíveis, com auxílio de séries temporais, métodos causais e redes neurais artificiais
(Centro Universitário do Estado do Pará, 2018) Teixeira, Rodrigo Simões; Siqueira, Sirius Raffael Jansen Costa; Freitas, Felipe Fonseca Tavares de; http://lattes.cnpq.br/5523511253031983; Nascimento, Polyana Santos Fonseca; http://lattes.cnpq.br/6889523334917369; Silva Junior, Carlos Gilberto Vieira da; http://lattes.cnpq.br/2738903947477853
The object of this study is demonstrate with three quantity ways about ideas of demand, what is: time series, casual method and artificial neural system. And with this results, organize comparations about routine data and find the better method to be apply in a gas station. Based on an error measurement, where the one presenting the lowest is considered the best method. In order to achieve this, the objective of this study is to evaluate which of the mathematical methods can be more efficient in forecasting demand. In analyses will be use Excel®, Crystal Ball® and Matlab®. After all work, it is possible see artificial neural system with the best results, in second place we have time series and for last casual method. When we consider the week analyses, it's verified the best results, because of data control and outliers finish. In this form, it is correct consider the artificial neural with the best method when the subject is ideas of demand.