77 results
Search Results
2. ESTUDO COMPARATIVO DO USO DE REDES NEURAIS ARTIFICIAIS E REGRESSÃO LINEAR MÚLTIPLA PARA A PREVISÃO DA CONCENTRAÇÃO CÁUSTICA EM UMA ETAPA DO PROCESSO DE FABRICAÇÃO DE ALUMINA.
- Author
-
Rozza, Giovanni Leopoldo, da Silva, Ruy Gomes, and Gama Müller, Sonia Isoldi Marty
- Abstract
Copyright of Revista Producao Online is the property of Associacao Brasileira de Engenharia de Producao and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
- Published
- 2015
- Full Text
- View/download PDF
3. INTELIGÊNCIA ARTIFICIAL EM GASTROENTEROLOGIA.
- Author
-
GOTARDELO, DAVI RIANI, SILVA, CAROLINNE LISBOA, ROSSONI, EMANUELA SINIMBU SILVA, DA ROSA, SARAH QUEIROZ, NAVES, GABRIEL RODRIGUES REZENDE, GOTARDELO, MARCELE PEREIRA SILVESTRE, and GOTARDELO, DANIEL RIANI
- Abstract
Copyright of Brazilian Journal of Surgery & Clinical Research is the property of Master Editora and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
- Published
- 2018
4. Modelo híbrido de previsão de séries temporais para possíveis aplicações no setor de geração eólica.
- Author
-
Verçosa Leal Junior, João Bosco, do Nascimento Camelo, Henrique, Lucio, Paulo Sérgio, and de Carvalho, Paulo Cesar Marques
- Abstract
Copyright of Revista Ciência e Natura is the property of Revista Ciencia e Natura and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
- Published
- 2018
- Full Text
- View/download PDF
5. APLICABILIDADE DO ALGORITMO DE LEVENBERG-MARQUARDT PARA ANÁLISE DE GERAÇÃO DE ENERGIA ELÉTRICA DE UM SISTEMA FOTOVOLTAICO.
- Author
-
Pinheiro, Elisângela, Rüther, Ricardo, and Lovato, Adalberto
- Abstract
Copyright of Revista Producao Online is the property of Associacao Brasileira de Engenharia de Producao and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
- Published
- 2017
- Full Text
- View/download PDF
6. APLICAÇÃO DE APRENDIZADO DE MÁQUINA PARA AUMENTO DE PRECISÃO DE UM SISTEMA AUTOMATIZADO DE NUTRIÇÃO DE SUÍNOS.
- Author
-
Dal Bo, Giuliano and Luis Corso, Leandro
- Subjects
SWINE nutrition ,ARTIFICIAL neural networks ,MACHINE learning ,SWINE farms ,ARTIFICIAL intelligence ,ANIMAL welfare ,COMPUTATIONAL neuroscience ,SOLID dosage forms - Abstract
Copyright of Revista Producao Online is the property of Associacao Brasileira de Engenharia de Producao and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
- Published
- 2022
7. APLICAÇÃO DE MODELOS DE PREVISÃO E INTELIGÊNCIA ARTIFICIAL PARA AVALIAR DEMANDA NO SEGMENTO DE SISTEMAS DE ILUMINAÇÃO AUTOMOTIVO.
- Author
-
Carrer Torres, José Luiz and Luís Corso, Leandro
- Abstract
Copyright of Tecno-Lógica is the property of Tecno-Logica and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
- Published
- 2021
- Full Text
- View/download PDF
8. REDE NEURAL ARTIFICIAL ARTMAP-FUZZY APLICADA NO RECONHECIMENTO DE FALHAS ESTRUTURAIS.
- Author
-
CHAVES, JACQUELINE S., LOPES, MARA LÚCIA M., ROBERTO CHAVARETTE, FÁBIO, and DOS ANJOS LIMA, FERNANDO PARRA
- Abstract
Copyright of Revista Iberoamericana de Ingeniería Mecánica is the property of Editorial UNED and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
- Published
- 2017
9. Dashboard Inteligente para apoio à tomada de decisão em empresa de courier.
- Author
-
Pinto Ferreira, Ricardo, Martiniano, Andréa, and Sassi, Renato José
- Abstract
Copyright of Revista Gestão & Tecnologia is the property of Revista Gestao & Tecnologia and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
- Published
- 2016
10. DESENVOLVIMENTO DE UM SISTEMA PARA AUXÍLIO AO DIAGNÓSTICO DE DIABETES EMPREGANDO REDES NEURAIS ARTIFICIAIS (SADD).
- Author
-
Vieira, João Paulo, Parreira, Fábio José, and Silveira, Sidnei Renato
- Abstract
Copyright of Computing & System Journal (C&S) / Revista de Sistemas e Computação (RSC) is the property of FACS Servicos Educacionais S.A. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
- Published
- 2016
11. DESEMPENHO DE TÉCNICAS E COMBINAÇÕES DE PREVISÕES: UM ESTUDO COM OS PERCENTUAIS RELACIONADOS COM O SISTEMA ÚNICO DE SAÚDE BRASILEIRO.
- Author
-
Werner, Liane and Martins, Vera
- Subjects
BUSINESS forecasting ,MEDICAL care ,REGRESSION analysis ,CUSTOMER services ,ARTIFICIAL neural networks - Abstract
Copyright of Revista Ingeniería Industrial is the property of Departamento de Ingenieria Industrial, Universidad del Bio-Bio and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
- Published
- 2015
12. Artificial Neural Network to analyse the customer profile of a sanitation company of São Paulo city.
- Author
-
Romero, M., Sassi, R. J., and Arrivabene, A.
- Subjects
ARTIFICIAL neural networks ,PRIVATE sector ,STOCKHOLDERS ,DATA marts ,STATISTICAL decision making - Abstract
The sanitation sector of São Paulo is responsible for a significant share of the brazilian market and the private sector participation in this segment has gained strength in recent years. To address the needs of the population and also give returns to shareholders, the sanitation company is constantly seeking to improve their results. Increase sales have become a major challenge for the managers of commercial areas. One way to increase the commercial efficiency is through the use of the large amount of information and data stored on corporate systems. The aim of this paper is to analyze the consumer profile of a sanitation company of São Paulo city, obtained from a Data Mart by a Multilayer Perceptron Artificial Neural Network. The results obtained are positive and demonstrate the acceptability of decision making [ABSTRACT FROM AUTHOR]
- Published
- 2012
13. ALGORITMO PARA O RECONHECIMENTO DE CARACTERES MANUSCRITOS.
- Author
-
Rocha Miranda, Rafael Arthur, da Silva, Francisco Assis, Pazoti, Mário Augusto, Artero, Almir Olivette, and Piteri, Marco Antonio
- Subjects
PATTERN recognition systems ,DIGITAL images ,COMPUTER vision ,NEURAL computers ,ARTIFICIAL neural networks ,OPTICAL character recognition - Abstract
Copyright of Colloquium Exactarum is the property of Asociacao Prudentina de Educacao e Cultura (APEC) and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
- Published
- 2013
- Full Text
- View/download PDF
14. Proposta de um Sistema Inteligente de Previsão de Colheita de Cogumelos.
- Author
-
Costa, Jorge, Branco, Frederico, Martins, José, Moreira, Fernando, Au-Yong-Oliveira, Manuel, Pérez-Cota, Manuel, González Castro, Miguel Ramón, and Díaz Rodríguez, María
- Abstract
Copyright of CISTI (Iberian Conference on Information Systems & Technologies / Conferência Ibérica de Sistemas e Tecnologias de Informação) Proceedings is the property of Conferencia Iberica de Sistemas Tecnologia de Informacao and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
- Published
- 2018
15. Simulação de moagem mista por rede neural artificial.
- Author
-
Rosa, Germano Mendes and da Luz, José Aurélio Medeiros
- Subjects
- *
ARTIFICIAL neural networks , *GRINDING machines , *SIMULATION methods & models , *PARTICLE size distribution , *DATA analysis , *MINERAL industries , *METALLURGICAL analysis , *STABILITY (Mechanics) - Abstract
This paper discusses the results of a mixed grinding simulator application based on an artificial neural network (multiple-layer perceptron using a back-propagation-like algorithm with moment). The data used came from a previous paper entitled "Selective grinding of dolomite and quartz mixes". The Shewhart control chart for individual values was used in order to verify the statistical stability of the simulation process results, which was useful for testing acceptance. The results have displayed good performance of this tool related to mix grinding simulation, a common issue in the mining and metallurgical sectors. [ABSTRACT FROM AUTHOR]
- Published
- 2012
- Full Text
- View/download PDF
16. ANÁLISE DE DADOS DE FISIOLOGIA DE PLANTAS APOIADA POR TÉCNICAS DE VISUALIZAÇÃO DE INFORMAÇÕES.
- Author
-
de Andrade Parisi, Lidiane, Artero, Almir Olivette, and Maia Souza, Gustavo
- Subjects
DATA visualization ,DATA analysis ,PLANT physiology ,TECHNOLOGICAL innovations ,ARTIFICIAL neural networks ,BAYESIAN analysis ,MATHEMATICAL models - Abstract
Copyright of Colloquium Exactarum is the property of Asociacao Prudentina de Educacao e Cultura (APEC) and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
- Published
- 2011
- Full Text
- View/download PDF
17. Coeficiente de rugosidade de Manning para o rio Paracatu.
- Author
-
Lyra, Guilherme B., Cecílio, Roberto A., Zanetti, Sidney S., and Lyra, Gustavo B.
- Subjects
SURFACE roughness ,ARTIFICIAL neural networks ,EXPANSION of solids ,GEOMETRY ,HYDRAULICS ,EQUATIONS ,SURFACES (Technology) ,RIVERS - Abstract
Copyright of Revista Brasileira de Engenharia Agricola e Ambiental - Agriambi is the property of Revista Brasileira de Engenharia Agricola e Ambiental and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
- Published
- 2010
- Full Text
- View/download PDF
18. Análise da trafegabilidade em estradas florestais utilizando métodos computacionais Analysis of the traffic performance in forest roads using computational methods
- Author
-
Robson Jose de Oliveira, José Marinaldo Gleriani, Carlos Cardoso Machado, Reginaldo Sérgio Pereira, and Sidney Araújo Cordeiro
- Subjects
Estradas não pavimentadas ,Transporte florestal ,Redes neurais artificiais ,Unpaved road ,Forest transportation ,Artificial neural networks ,Forestry ,SD1-669.5 - Abstract
As estradas florestais são o principal meio de integração entre as florestas e as empresas. A partir do exposto, percebe-se a necessidade não apenas da correta aplicação de atividades de manutenção, mas também de se determinar o tempo exato para tal intervenção. Partindo desse pressuposto, este trabalho apresenta os resultados da apreciação de dois métodos de classificação da qualidade de estradas não pavimentadas, com o intuito de se verificar a aplicabilidade dos mesmos na caracterização das estradas florestais brasileiras e servir como base para um sistema de gestão das operações de manutenção destas vias. Foram medidos os principais defeitos em estradas florestais seguindo um método denominado de Índice de Condição de Rodovia Não Pavimentada (ICRNP), que serviram de base para gerar um banco de dados para testar a eficiência do uso de redes neurais artificiais (RNA) na administração das estradas florestais, minimizando custos e paralisações de tráfego. Concluiu-se que a utilização das redes neurais artificiais apresentou resultados superiores ao método do ICRNP.The forest roads are the main integration mode between forests and companies. Therefore, there is a need for defining not only the required maintenance activities, but also the exact time for such intervention. Starting from this premise this paper presents the results of the assessment of two methods of classification of the quality of unpaved roads in order to verify which one reflects the field reality and thus can serve as the basis for a unpaved road management system. In this paper main defects in forest roads were measured following a method named Unsurfaced Road Condition Index (URCI) which served as a database for testing the efficiency of using artificial neural networks (ANN) in the management of forest roads taking into account costs minimization and traffic stoppage. It was concluded that the use of artificial neural networks showed superior performance that the URCI method.
- Published
- 2013
19. USO DE NEUROSIMULACIÓN EN EL POSICIONAMIENTO DE POZOS DURANTE EL DESARROLLO DE UN CAMPO DE HIDROCARBUROS
- Author
-
Elkin Rodolfo Santafe Rangel and Hector Emiliano Barrios Molano
- Subjects
Neurosimulation ,artificial neural networks ,reservoirs numeric simulation ,mature fields ,hydrocarbon field development ,free software ,python ,Electrical engineering. Electronics. Nuclear engineering ,TK1-9971 ,Renewable energy sources ,TJ807-830 - Abstract
This paper shows the implementation of a neurosimulation technique for the well placement applied to the development of a heterogeneous hydrocarbon field with an irregular geometry. During the development of a hydrocarbon field the well placement is a major task, because a small change in location can make gains or losses of money during the remaining productive life of the field; this paper presents a neurosimulation technique as an alternative to conventional methods of well placement which are expensive and consume large amounts of time. This technique is a bridge between hard-computing and soft-computing; effectively mixes artificial neural networks (ANN) and numerical reservoir simulation, in this way using the numerical reservoir simulation in a combination of training wells, production data are obtained along with other data which are used to train and adjust the network, then a large number of scenarios are generated which are evaluated by the trained ANN, the best results are verified whit the numerical reservoir simulation, and then it is possible to predict the rate at which the wells will produce and the cumulative hydrocarbon production. For the development of this work open source tools and free software was used to encourage their use and development in research, in academia and in hydrocarbon industry. This work shows an alternative method of selecting wells that produce fast and accurate results, with which it is easy to take the decision about where is the best place to drill new wells during the development of a hydrocarbon field.
- Published
- 2009
20. Segmentação do mercado consumidor cativo alta e média tensão de uma distribuidora de energia elétrica: aplicação do Mapa Auto-Organizável de Kohonen para descoberta de padrões de inadimplência no setor.
- Author
-
da Silva Carvalho, Norma Alice, Castro Souza, Reinaldo, and Kahn Epprecht, Eugenio
- Abstract
Copyright of Exacta is the property of Exacta - Engenharia de Producao and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
- Published
- 2016
- Full Text
- View/download PDF
21. Disaggregated approach to urban trip distribution: a comparative analysis between artificial neural networks and discrete choice models
- Author
-
Marina Urano de Carvalho Caldas, Cira Souza Pitombo, Felipe Lobo Umbelino de Souza, and Renan Favero
- Subjects
Destination choices ,Multinomial Logit ,Nested Logit ,Artificial Neural Networks ,Transportation engineering ,TA1001-1280 - Abstract
Discrete choice models have been used over the years in disaggregated approaches to forecast destination choices. However, there are important constraints in some of these models that pose obstacles to using them, such as the Independence of Irrelevant Alternatives (IIA) property in the Multinomial Logit model, the need to assume specific structures and high calibration times, depending on the complexity of the case being evaluated. However, some of these mentioned constraints could be mitigated using Mixed Models or Nested Logit. Therefore, this paper proposes a comparative analysis between the Artificial Neural Network (ANNs), the Multinomial and Nested Logit models for disaggregated forecasting of urban trip distribution. A case study was conducted in a medium-sized Brazilian city, Santa Maria (RS), Brazil. The data used come from a household survey, prepared for the Urban Mobility Master Plan. For the sake of comparison, hit rates and frequency of trip distribution distances were analyzed, showing that ANNs can be as efficient as the Discrete Choice models for disaggregated forecasting of urban trip destination without, however, assuming some constraints. Finally, based on the results obtained, the efficiency of ANNs is observed for predicting alternatives with a low number of observations. They are important tools for obtaining Origin-Destination matrices from incomplete sample matrices or with a low number of observations. However, it is important to mention that discrete choice models can provide important information for the analyst, such as statistical significance of parameters, elasticities, subjective value of attributes, etc.
- Published
- 2022
- Full Text
- View/download PDF
22. Prótese mioelétrica controlada por redes neurais.
- Author
-
de Godoi, Tomás da Silva Martins, Duque, Luciano Henrique, and de Obaldía Díaz, Francisco Javier
- Abstract
Copyright of Universitas. Gestão e Tecnologia is the property of Centro Universitario de Brasilia, UNICEUB and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
- Published
- 2013
- Full Text
- View/download PDF
23. Quanto pior, melhor: Estudo da utilização da análise por envoltória de dados em modelos de análise de inadimplência/insolvência de empresas.
- Author
-
Pereira de Castro Casa Nova, Silvia
- Subjects
- *
BANKRUPTCY , *DATA envelopment analysis , *ACCOUNTING , *FINANCIAL statements , *LOGISTIC regression analysis , *ARTIFICIAL neural networks - Abstract
This paper aims to develop an insolvency/failure analysis model allying accounting information with Data Envelopment Analysis (DEA). The research was conducted in phases: initially previous studies were examined; then samples were obtained from different databases containing financial information of distressed and healthy companies; and, finally a methodology for application of DEA to failure/insolvency analysis was proposed. The results were comparable to those obtained from logistic regression models and neural network models, without, however, a clear indication of the superiority of any procedure. [ABSTRACT FROM AUTHOR]
- Published
- 2013
- Full Text
- View/download PDF
24. Formulação para tensão de flambagem distorcional em colunas com seção U enrijecido de chapa de aço dobrada a frio.
- Author
-
Santos, Fellipe Peixoto, de Jesus Nagahama, Koji, and de Souza Matos Gadéa, Anderson
- Subjects
- *
FINITE strip method , *MECHANICAL buckling , *DEFORMATIONS (Mechanics) , *COLD-formed steel , *NUMERICAL analysis , *ARTIFICIAL neural networks - Abstract
This paper presents a formulation for distortional buckling stress (σdist) in cold-formed steel lipped U section columns, with simply supported end conditions and free warping. This formulation is based on an adjusted model from an artificial neural network (ANN), whose data was provided by a program based on the Finite Strip Method (FSM). The results show the viability of the obtained equation for determining σdist on a columns made with the analyzed sections. [ABSTRACT FROM AUTHOR]
- Published
- 2013
- Full Text
- View/download PDF
25. UTILIZAÇÃO DE REDES NEURAIS ARTIFICIAIS COMO ESTRATÉGIA DE PREVISÃO DE PREÇOS NO CONTEXTO DE AGRONEGÓCIO.
- Author
-
Ferreira, Luciano, de Moura, Gilnei Luiz, Borenstein, Denis, and Fischmann, Adalberto Américo
- Subjects
ARTIFICIAL neural networks ,AGRICULTURAL industries ,FARM produce prices ,BUSINESS forecasting ,STRATEGIC planning ,DECISION making ,COGNITIVE bias - Abstract
Copyright of Revista de Administração e Inovação (RAI) is the property of Revista de Administracao e Inovacao- RAI and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
- Published
- 2011
- Full Text
- View/download PDF
26. REDES NEURAIS ARTIFICIAIS E SEGMENTAÇÃO PSICOGRÁFICA EM MARKETING.
- Author
-
Omaki, Eduardo Tadayoshi, Fonseca, Francisco Ricardo Bezerra, and de Mello, Sérgio Carvalho Benício
- Subjects
- *
MARKETING , *MARKET segmentation , *ARTIFICIAL neural networks , *PURCHASING agents , *PSYCHOGRAPHICS , *RISK perception , *INDUSTRIAL marketing - Abstract
The dynamics and complexity of contemporary markets have made it difcult for marketing professionals to understand the buying behavior of industrial decision-makers. Efcient market segmentation can improve marketing strategies and knowledge of buyers, favoring a match between offer and demand. This paper discusses the possibility of developing an Articial Neural Network (ANN) to identify and build a prole of the professional purchasing agent, by means of a psychographic model based on perceived risk. The construction of this network is based on the Perceptron Multi-Layer model and the Backpropagation learning algorithm. This type of network is capable of identifying data patterns, which do not show linear characteristics, a common feature of psychographic data. This technology can reveal patterns and develop skills that enable a better understanding of industrial markets. [ABSTRACT FROM AUTHOR]
- Published
- 2010
27. USO DE INTELIGÊNCIA COMPUTACIONAL NO DESENVOLVIMENTO, ADEQUAÇÃO E CONTROLE DE QUALIDADE DE AÇOS LAMINADOS A QUENTE NA USIMINAS-CUBATÃO.
- Author
-
de Morais, Willy Ank and Borges, Herbert Christian
- Subjects
- *
ARTIFICIAL intelligence , *COMPUTATIONAL intelligence , *ARTIFICIAL neural networks , *QUALITY control , *STEEL industry , *PRODUCT quality - Abstract
The Technical Assistance and Integrated Control of Usiminas-Cubatão - have studied and analyzed the microstructural characteristics and mechanical properties of the structural steels the company produces as well as the phenomenological models it has created to describe their properties. An important tool that uses artificial intelligence was recently incorporated: the Adaptive Neural Networks (ANN's). This paper shows how the ANN's are improving the product development activities, as well as the adequacy and application of hot rolled products. The chemical composition, the rolling temperatures, the final product size and position of sampling are parameters used in the ANN's, making it possible to select effectively and practically, and not only theoretically, the best combinations of these parameters to obtain a more appropriate product. The application of products has been benefited by the use of these tools, especially for petrochemical and automobile uses, as well as the internal activities of product quality control. [ABSTRACT FROM AUTHOR]
- Published
- 2010
- Full Text
- View/download PDF
28. O IMPACTO DA TAXA DE CÂMBIO NO APREÇAMENTO DE OPÇÕES NO BRASIL -- UMA ANÁLISE COMPARATIVA ENTRE UM MODELO DE REDE NEURAL E O MODELO DE BLACK & SCHOLES.
- Author
-
de Planas, Carlos Alberto Aragón, Ferreira, Léo da Rocha, and Lachtermacher, Gerson
- Subjects
- *
ECONOMIC impact , *FOREIGN exchange rates , *PRICING , *COMPARATIVE economics , *ARTIFICIAL neural networks , *ECONOMIC models , *ECONOMIC forecasting - Abstract
The main goal of this paper is to evaluate the impact of the exchange rate volatility in price prediction of derivative securities in the Brazilian capital markets using an artificial neural network technique, given the Black & Scholes Model limitations. For this purpose a multiplayer, feedforward neural network, trained by the backpropagation algorithm model, to perform the prediction of the Telemar option prices was developed. The model results show that price estimates are close to the real values, mainly when appended to the exchange rate, confirming that the performance of neural network is superior to other results. The inclusion of the exchange rate in neural networks technique results in a better price forecasting for the options, because the volatility in the price of the underlying asset is caused by temporary arbitrage of quotes among the national and foreign stock markets where the company is listed. [ABSTRACT FROM AUTHOR]
- Published
- 2009
29. Redes neurais artificiais e padrões de falência.
- Author
-
Thompson, Carlos A., Caetano, Fábio M., and Thompson, Marly F.
- Subjects
FINANCIAL risk management ,ARTIFICIAL neural networks ,ALGORITHMS ,FINANCIAL performance ,BANKRUPTCY ,DEFAULT (Finance) ,LIQUIDITY (Economics) ,PROFITABILITY ,ECONOMIC activity - Abstract
Copyright of Revista de Economia e Administração is the property of INSPER Instituto de Ensino e Pesquisa and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
- Published
- 2004
- Full Text
- View/download PDF
30. Estimativa de valores ausentes com redes neurais artifciais: o caso dos custos de construção civil.
- Author
-
Neto, Luiz Biondi, de Mello, João Carlos Correia Baptista Soares, Gomes, Eliane Gonçalves, and Meza, Lidia Angulo
- Subjects
REAL property sales & prices ,FINANCE ,CONSTRUCTION cost estimates ,SELLING ,ARTIFICIAL neural networks ,COMPARATIVE studies ,REAL estate business ,ARTIFICIAL intelligence - Abstract
Copyright of Revista de Economia e Administração is the property of INSPER Instituto de Ensino e Pesquisa and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
- Published
- 2004
- Full Text
- View/download PDF
31. MODELOS DE MACHINE LEARNING PARA PREDIÇÃO DO SUCESSO DE STARTUPS.
- Author
-
Rodrigues, Fabiano, Aparecido Rodrigues, Francisco, and Rocha Rodrigues, Thelma Valéria
- Subjects
MACHINE learning ,NEW business enterprises ,GOING public (Securities) ,INVESTMENTS ,SUPPORT vector machines ,ARTIFICIAL neural networks ,LOGISTIC regression analysis - Abstract
Copyright of Journal of Business & Projects / Revista de Gestão e Projetos is the property of Revista de Gestao e Projetos and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
- Published
- 2021
- Full Text
- View/download PDF
32. Artificial neural networks application to predict bond steel-concrete in pull-out tests
- Author
-
Lorenzi, Alexandre, Silva, Bruno do Vale, Barbosa, Mônica Pinto, and Silva Filho, Luiz Carlos Pinto da
- Subjects
Redes neurais artificiais ,Artificial neural networks ,Pull-out test ,Bond steel-concrete ,APULOT test ,Aderência aço-concreto ,Concreto ,Concrete strength ,Resistência à compressão - Abstract
O estudo visa avaliar a possibilidade de se usar os resultados do ensaio de arrancamento “pull-out test” – ensaio de aderência aço-concreto para estimativa da resistência à compressão do concreto, este método vem sendo utilizado com sucesso pelo grupo de pesquisa APULOT, desde 2008 [1]. A pesquisa ora realizada evidencia a existência da correlação entre essas duas variáveis, aderência e resistência à compressão do concreto, o que permite determinar estimativas apropriadas da resistência à compressão do concreto, melhorando deste modo a capacidade do controle tecnológico “in situ” do concreto. Entretanto para se obter respostas adequadas dos ensaios de aderência aço-concreto é necessário controlar as configurações de ensaio, dado que existem diversos formatos de corpos de prova para estes tipos de ensaios na literatura. Deste modo, este trabalho tem por objetivo correlacionar os resultados obtidos em ensaios de aderência do tipo pull-out a suas variáveis por meio da utilização de Redes Neurais Artificiais (RNA). Com a utilização de uma RNA, pode-se correlacionar, de forma não linear, dados de entrada conhecidos (idade de ruptura, comprimento de ancoragem, cobrimento e resistência à compressão) com parâmetros de controle (tensão de aderência aço-concreto). Para gerar o modelo neural é necessário treinar a rede, expondo-a a uma série de dados com parâmetros de entrada e de saída conhecidos. Isto permite estimar os coeficientes de correlação entre os neurônios de cada camada. O presente trabalho apresenta a modelagem de uma RNA capaz de realizar uma aproximação não linear, visando estimar a resistência à compressão do concreto a partir dos resultados de ensaios de aderência aço-concreto. This study aims the possibility of using the pull-out test results – bond tests steel-concrete, that has been successfully carried out by the research group APULOT since 2008 [1]. This research demonstrates that the correlation between bond stress and concrete compressive strength allows estimate concrete compressive strength. However to obtain adequate answers testing of bond steel-concrete is necessary to control the settings test. This paper aims to correlate the results of bond tests of type pull-out with its variables by using Artificial Neural Networks (ANN). Though an ANN is possible to correlate the known input data (age rupture, anchorage length, covering and compressive strength of concrete) with control parameters (bond stress steel-concrete). To generate the model it is necessary to train the neural network using a database with known input and output parameters. This allows estimating the correlation between the neurons in each layer. This paper shows the modeling of an ANN capable of performing a nonlinear approach to estimate the concrete compressive strength using the results of steel-concrete bond tests.
- Published
- 2017
33. Autoproteção para a internet das coisas
- Author
-
Almeida, Fernando Mendonça de, Ribeiro, Admilson de Ribamar Lima, and Moreno, Edward David
- Subjects
Redes neurais artificiais ,Internet ,Algoritmo de células dendríticas ,Internet of things ,Algoritmos de computador ,Autoproteção ,Artificial neural networks ,Redes neurais (Computação) ,Dendritic cells algorithm ,Internet das coisas ,Self-protection ,Computação ,CIENCIA DA COMPUTACAO [CIENCIAS EXATAS E DA TERRA] - Abstract
Fundação de Apoio a Pesquisa e à Inovação Tecnológica do Estado de Sergipe - FAPITEC/SE The Internet of Things is a new paradigm of communication based on the ubiquitous presence of objects that, having unique address, they can cooperate with their peers to achieve a common goal. Applications in several areas can benefit from this new paradigm, but the Internet of Things is very vulnerable to attack. The large number of connected devices make an autonomic approach necessary and the small amount of resources requires the use of efficient techniques. This paper proposes a self-protection architecture for the Internet of Things using Artificial Neural Network and Dendritic Cells Algorithm, two bio-inspired techniques. The experiments of this paper show that the use of these two techniques is possible. The Artificial Neural Network implementation consume a small memory footprint, having a high accuracy rate and the Dendritic Cells Algorithm show to be interesting for it distributivity, allowing better use of network resources. A Internet das Coisas é um novo paradigma de comunicação baseado na presença ubíqua de objetos que, através de endereçamento único, cooperam com seus pares para atingir um objetivo em comum. Aplicações em diversas áreas podem se beneficiar dos conceitos da Internet das Coisas, porém esta rede é muito vulnerável a ataques, seja pela possibilidade de ataque físico, pela alta conectividade dos dispositivos, a enorme quantidade de dispositivos conectados ou a baixa quantidade de recursos disponíveis. A grande quantidade de dispositivos conectados faz com que abordagens autonômicas sejam necessárias e a reduzida quantidade de recursos exige a utilização de técnicas eficientes. Este trabalho propõe uma arquitetura de autoproteção para a Internet das Coisas utilizando as técnicas de Rede Neural Artificial e Algoritmo de Células Dendríticas, duas técnicas bio-inspiradas que, através de experimentos, mostraram a possibilidade de serem utilizadas na Internet das Coisas. A implementação da Rede Neural Artificial utilizada consumiu poucos recursos de memória do dispositivo, mantendo uma alta taxa de acerto, comparável a trabalhos correlatos que não se preocuparam com o consumo de recursos. A utilização do Algoritmo de Células Dendríticas se mostrou interessante pela sua distributividade, permitindo uma melhor utilização dos recursos da rede, como um todo.
- Published
- 2016
34. Análise de risco de crédito apoiada por inteligencia artificial
- Author
-
Silva, Renato Aparecido Souza da, Fávero, Patrícia Belfiore, Kurashima, Celso Setsuo, and Leonardi, Fabrizio
- Subjects
CREDITY RISK ANALISYS ,PROGRAMA DE PÓS-GRADUAÇÃO EM ENGENHARIA DA INFORMAÇÃO - UFABC ,ANÁLISE DE RISCO ,ARTIFICIAL NEURAL NETWORKS ,ARTIFICIAL INTELLIGENCE ,REDES NEURAIS ,INTELIGÊNCIA ARTIFICIAL - Abstract
Orientadora: Profa. Dra. Patricia Belfiore Fávero Dissertação (mestrado) - Universidade Federal do ABC, Programa de Pós-Graduação em Engenharia da Informação, 2016. Este trabalho tem o objetivo de analisar as conjunturas do mercado financeiro, no tocante ao assunto de avaliação de risco de crédito. Esta atividade é primordial em toda e qualquer instituição que trabalhe com a concessão de ativos de crédito a credores, assumindo com isso todos os riscos de perda financeira envolvidos em uma operação de confiança. Com isso, este trabalho apresenta diferentes métodos de avaliação de risco de crédito utilizados por instituições ao redor do mundo, destacando melhores práticas de avaliação dos riscos na concessão de crédito, e características locais especificas que muitas vezes são similares em muitos países como regulamentações dos órgãos reguladores do governo local, e até mesmo acordos internacionais de melhores práticas de gestão de carteira de crédito como a Basileia. Dentre os métodos mais comumente utilizados nas instituições, este trabalho presta uma avaliação mais próxima sobre os métodos estatísticos aplicados para avaliação de risco de crédito, e os avanços conquistados com o apoio de técnicas sofisticadas de inteligência artificial no auxílio a resolução de problemas de análise de crédito, como redes neurais artificiais, algoritmos genéticos, e aprendizado conjunto de máquina. A união entre os conhecimentos clássicos de gestão de crédito e as técnicas de inteligência artificial permitem uma melhor assertividade na predição de bons e maus clientes, garantindo um menor risco financeiro e uma economia considerável devido à redução de processos operacionais. Com base nas análises literárias analisadas, será apresentado um estudo experimental, com o objetivo de apresentar na pratica o uso de técnicas de inteligência artificial para simular a resolução de problemas simples de análise de crédito. This work has been developed with the objective of analyzing different situations in the financial Market, as regards the issue of credit risk evaluation. This activity is paramount in any institution that works with the granting of credit assets to creditors, taking with it all the risks of financial loss involved in a reliable operation. Thus, this paper presents various credit risk assessment methods used by institutions around the world, highlighting best risk assessment practices in lending, and specific local characteristics that are often similar in many countries such regulations organs regulators of local government, and even international agreements to better credit portfolio management practices such as Basel. Among the methods most commonly used in institutions, this paper provides a closer assessment of the statistical methods used to evaluate credit risk, and the advances made with the support of sophisticated artificial intelligence techniques to aid solving analysis problems credit, such as artificial neural networks, genetic algorithms, and joint learning machine. The union between the classical knowledge credit management and artificial intelligence techniques, provide better assertiveness in the prediction of good and bad clients, ensuring a lower financial risk and considerable savings due to reduced operational processes. Based on the literary analysis analyzed, an experimental study will be presented, with the aim of presenting in practice the use of artificial intelligence techniques to simulate the resolution of simple credit analysis problems.
- Published
- 2016
35. Intelligent system for improving dosage control
- Author
-
Fabio Cosme Rodrigues dos Santos, André Felipe Henriques Librantz, Cleber Gustavo Dias, and Sheila Gozzo Rodrigues
- Subjects
water treatment plant ,process control ,coagulant dosage ,artificial neural networks ,optimization. ,Engineering (General). Civil engineering (General) ,TA1-2040 ,Science (General) ,Q1-390 - Abstract
Coagulation is one of the most important processes in a drinking-water treatment plant, and it is applied to destabilize impurities in water for the subsequent flocculation stage. Several techniques are currently used in the water industry to determine the best dosage of the coagulant, such as the jar-test method, zeta potential measurements, artificial intelligence methods, comprising neural networks, fuzzy and expert systems, and the combination of the above-mentioned techniques to help operators and engineers in the water treatment process. Current paper presents an artificial neural network approach to evaluate optimum coagulant dosage for various scenarios in raw water quality, using parameters such as raw water color, raw water turbidity, clarified and filtered water turbidity and a calculated Dose Rate to provide the best performance in the filtration process. Another feature in current approach is the use of a backpropagation neural network method to estimate the best coagulant dosage simultaneously at two points of the water treatment plant. Simulation results were compared to the current dosage rate and showed that the proposed system may reduce costs of raw material in water treatment plant.
- Published
- 2017
- Full Text
- View/download PDF
36. Forecasts for the Canadian Lynx time series using method that bombine neural networks, wavelet shrinkage and decomposition
- Author
-
Levi Lopes Teixeira, Paulo Henrique Siqueira, Luiz Albino Teixeira Jr, Samuel Bellido Rodrigues, and Arinei Carlos Lindbeck da Silva
- Subjects
Wavelet Shrinkage ,Wavelet Decomposition ,Artificial neural networks ,Industrial engineering. Management engineering ,T55.4-60.8 - Abstract
Time series forecasting is widely used in various areas of human knowledge, especially in the planning and strategic direction of companies. The success of this task depends on the forecasting techniques applied. In this paper, a hybrid approach to project time series is suggested. To validate the methodology, a time series already modeled by other authors was chosen, allowing the comparison of results. The proposed methodology includes the following techniques: wavelet shrinkage, wavelet decomposition at level r, and artificial neural networks (ANN). Firstly, a time series to be forecasted is submitted to the proposed wavelet filtering method, which decomposes it to components of trend and linear residue. Then, both are decomposed via level r wavelet decomposition, generating r + 1 Wavelet Components (WCs) for each one; and then each WC is individually modeled by an ANN. Finally, the predictions for all WCs are linearly combined, producing forecasts to the underlying time series. For evaluating purposes, the time series of Canadian Lynx has been used, and all results achieved by the proposed method were better than others in existing literature.
- Published
- 2015
- Full Text
- View/download PDF
37. Predicting Movement of Homeless Young Adults: Artificial Neural Networks and Generalized Linear Models.
- Author
-
Helderop, Edward, Ferguson, Kristin M., Grubesic, Tony H., and Bender, Kimberly
- Subjects
REGRESSION analysis ,EMPLOYMENT of young adults ,LINEAR models (Communication) ,HOMELESS persons ,ARTIFICIAL neural networks ,PSYCHOLOGY - Abstract
Objective: Previous research has indicated high rates of intercity movement among homeless young adults (HYAs) for a variety of prosocial (e.g., avoiding domestic violence and seeking new employment opportunities) and antisocial (e.g., following drug supplies and avoiding law enforcement) reasons. The complicated mixture of individual circumstances associated with transience has made it difficult to predict features of transience, such as distance traveled and move frequency. Method: This study describes a method to build an artificial neural network (ANN) that predicts distance traveled and compares the results of thatANN to a generalized linear regression. Results: The ANN more accurately predicts distance traveled than does the linear statistical model and advances the development of approaches to predict complicated human phenomena. Conclusions: Accurately predicting features of transience among HYAs is important in tailoring effective interventions aimed at minimizing travel for negative reasons and making travel for positive reasons safer. [ABSTRACT FROM AUTHOR]
- Published
- 2018
- Full Text
- View/download PDF
38. Estimativa de alturas geoidais para o estado de São Paulo baseada em redes neurais artificais
- Author
-
Veronez, Maurício Roberto, Souza, Sergio Florencio de, Matsuoka, Marcelo Tomio, and Reinhardt, Alessandro Ott
- Subjects
Redes neurais artificiais ,Artificial neural networks ,EGM2008 ,GPS ,MAPGEO2004 ,Geoidal height - Abstract
A informação da altitude fornecida pelo sistema GNSS (Global Navigation Satellite System) é puramente geométrica, e na maioria dos trabalhos de engenharia a altitude deve estar referenciada ao geóide. Com um número suficiente de Referências de nível (Rn’s) com coordenadas horizontais e verticais conhecidas, quase sempre, e´possível ajustar-se, pelo Método dos Mínimos Quadrados, expressões matemáticas que permitem interpolar as alturas geoidais. O objetivo deste trabalho foi avaliar a eficiência das Redes Neurais Artificiais (RNAs) no processo de predição de alturas geoidais tendo como área de estudo o Estado de São Paulo. As informações utilizadas basearam-se em um conjunto de 157 Referências de nível (Rn’s) distribuídas uniformemente em todo Estado. Para estas Rn’s são conhecidas suas coordenadas horizontais (latitude e longitude) e verticais (altitudes geométrica e ortométrica e altura geoidal). Das 157 Rn’s, 115 foram utilizadas para o treinamento da RNA e 42 no processo de simulação para avaliar a eficiência do modelo proposto. A eficiência baseou-se em determinar as discrepâncias (erro) entres as alturas geoidais conhecidas e as obtidas pelo modelo neural. Como contribuição da pesquisa comparou-se também os valores simulados com o Earth Gravitational Model 2008 (EGM2008) e também com o MAPGEO2004. Em termos de resultados a RNA proporcionou um erro absoluto médio de 0,19 m ±0,14 m com uma forte correlação (R2 = 0,9871) com os valores tomados como verdadeiros. Estatisticamente os testes realizados mostraram que não houve diferença entre as médias das alturas geoidais conhecidas e as fornecidas pelo modelo neural para um nível de significância de 5%. Comparando-se os resultados com o EGM2008 e MAPGEO2004 a RNA proporcionou uma redução no erro de 0,07 m e 0,44 m, respectivamente. The information of height provided by the GNSS (Global Navigation Satellite System) is purely geometrical, and in most engineering papers, the height must be referenced to the geoid. Provided we have a sufficient number of Bench Marks (BMs) with known horizontal and vertical coordinates, it is nearly always possible to adjust mathematical expressions that allow for the interpolation of geoidal heights. The aim of this paper is to evaluate the efficiency of Artificial Neural Network (ANN) in the process of predicting geoidal heights, having the State of São Paulo as the area of study. The information used is based on a set of 157 BMs, evenly distributed all across the State. The horizontal coordinates (latitude and longitude) and the vertical coordinates (geometrical, orthometrical and geoidal heights) of these BMs are known. From the 157 BMs, 115 were used for the training of RNA and 42 in the process of simulation to assess the efficiency of the model proposed. Efficiency is based in determining the discrepancies (error) between known geoidal heights and those which were obtained by the neural model. As a contribution to this research, we have compared the values simulated with the Earth Gravitational Model 2008 (EGM2008) and with the MAPGEO2004 as well. In terms of results, the RNA produced a mean absolute error of 0.19 m ± 0.14 m and a strong correlation ( R2 = 0.9871) with the values taken as true. Statistically, the tests showed that there was no difference between known geoidal heights and those which were provided by the neural model for a level of significance of 5%. When we compare these results with the EGM2008 and MAPGEO2004, the RNA has an error reduction of 0.07 and 0.44 m, respectively.
- Published
- 2009
39. Identificación de suelos expansivos y colapsables en el noreste de Brasil a partir de Redes Neuronales Artificiales generadas en Pernambuco
- Author
-
Holanda, Maria Julia de Oliveira, Ferreira, Silvio Romero de Melo, Amorim, Samuel Franca, Borges, Jesce John Silva, and Silva, Larissa Ferreira da
- Subjects
Collapsible Soils ,Classification equation ,Solos Colapsíveis ,Suelos Expansivos ,Suelos Colapsables ,Expansive Soils ,Redes Neuronales Artificiales ,Ecuación de clasificación ,Equação de classificação ,Solos Expansivos ,Artificial Neural Networks ,Redes Neurais Artificiais - Abstract
Collapsible and expansive soils are problematic in Civil Engineering, causing pathologies in buildings due to the variation in volume with the change in humidity. The identification of these soils in the design phase is important. The paper aims to develop an Artificial Neural Network architecture trained with soils from Pernambuco, to identify expansive and collapsible soils, and expand its application to soils from other states in Northeastern Brazil. Developed from 87 samples, divided between training (53 samples), selection (17 samples) and test (17 samples) groups, according to 4 input variables, percentage of sand, percentage of clay, plasticity and activity indices. The best network architecture consists of 4 neurons at the input and 1 at the output. For the blind validation of the model, the network was applied to 45 samples of collapsible and expansive soils from other Northeastern states. The performance analysis of the classification accuracy of the network with data from Pernambuco showed an accuracy rate of 76.5% and in the validation in the other Northeastern states, pattern recognition was even higher, reaching an accuracy of 91.1%, demonstrating capacity capturing trends in soil surface movement and aiding in problem solving. Los suelos colapsables y expansivos son problemáticos en la Ingeniería Civil, provocando patologías en los edificios debido a la variación de volumen con el cambio de humedad. La identificación de estos suelos en la fase de diseño es importante. El artículo tiene como objetivo desarrollar una arquitectura de Red Neural Artificial entrenada con suelos de Pernambuco, para identificar suelos expansivos y colapsables, y ampliar su aplicación a suelos de otros estados del noreste de Brasil. Desarrollado a partir de 87 muestras, divididas entre los grupos de entrenamiento (53 muestras), selección (17 muestras) y prueba (17 muestras), según 4 variables de entrada, porcentaje de arena, porcentaje de arcilla, índices de plasticidad y actividad. La mejor arquitectura de red consta de 4 neuronas en la entrada y 1 en la salida. Para la validación ciega del modelo, se aplicó la red a 45 muestras de suelos colapsables y expansivos de otros estados del Noreste. El análisis de desempeño de la precisión de clasificación de la red con datos de Pernambuco arrojó una tasa de precisión del 76.5% y en la validación en los otros estados del Noreste, el reconocimiento de patrones fue aún mayor, alcanzando una precisión del 91.1%, demostrando capacidad de captura de tendencias en el movimiento de la superficie del suelo y ayudar en la resolución de problemas. Solos colapsíveis e expansivos são problemáticos na Engenharia Civil causando patologias nas edificações devido à variação de volume com a mudança de umidade. A identificação desses solos na fase de projeto é importante. O artigo visa elaborar uma arquitetura de Rede Neural Artificial treinada com solos de Pernambuco, para identificação de solos expansivos e colapsíveis, e ampliar sua aplicação à solos dos demais estados do Nordeste brasileiro. Desenvolvida a partir de 87 amostras, divididas entre grupos de treinamento (53 amostras), seleção (17 amostras) e teste (17 amostras), segundo 4 variáveis de entrada porcentagem de areia, porcentagem de argila, índices de plasticidade e atividade. A melhor arquitetura da rede consiste em 4 neurônios na entrada e 1 na saída. Para a validação às cegas do modelo, a rede foi aplicada a 45 amostras de solos colapsíveis e expansivos de demais estados do Nordeste. A análise de desempenho da precisão de classificação da rede com dados de Pernambuco apresentou uma taxa de acurácia de 76,5% e na validação nos demais estados do Nordeste o reconhecimento de padrões foi ainda maior, atingindo acurácia de 91,1%, demonstrando capacidade de capturar tendências no movimento da superfície do solo e auxiliando na resolução de problemas.
- Published
- 2021
40. Estimación de la resistencia a la penetración de suelos usando redes neuronales artificiales Prediction of the soils penetration strength using artificial neural networks
- Author
-
Nidia Johana Valdés Holguín, Luis Octavio González Salcedo, and Adrián L E. Will
- Subjects
Suelo ,compactación del suelo ,inteligencia artificial ,redes neuronales artificiales ,resistencia a la penetración ,Artificial intelligence ,artificial neural networks ,soils ,soil compaction ,soil penetration strength ,Agriculture - Abstract
Las redes neuronales artificiales, simuladoras del proceso de aprendizaje de las neuronas biológicas, han sido utilizadas con éxito en el cálculo de parámetros en diversos problemas de ingeniería en que las variables involucradas tienen una alta relación no lineal entre sí y la modelación no permite representar el problema mediante una función matemática de fácil deducción. En la ciencia del suelo la predicción de algunas propiedades involucra diversas variables que hacen de su estimación por medio de modelos matemáticos un proceso complejo, y trasladan la solución del problema al campo de la inteligencia artificial. En el presente artículo se reporta la elaboración de redes neuronales artificiales para la estimación de la resistencia a la penetración a diferentes profundidades de un suelo; se consideran como variables influyentes el contenido de humedad, la densidad, la carga estática y la presión de inflado. Los resultados muestran una mejor estimación para profundidades entre 20 cm y 30 cm.Artificial Neural Networks simulate the learning process of biological neurons, and these have been successfully used in the computation of parameters on several engineering problems where exist a strong nonlinear relation among the variables. In soil science, estimation of some properties involves variables that are complicated to estimate using mathematical models, so the solution for the problems fall into the field of Artificial Intelligence. The present paper reports the elaboration of an Artificial Neural Network for the estimation of penetration resistance of soil at different depths, considering as influential variables humidity, density, static load, and inflate pressure. The best estimation results were obtained at a depth of 20-30 cm.
- Published
- 2011
41. A preparação de material terminológico em língua inglesa por meio de ferramentas linguístico-computacionais Preparation of terminological material in english by means of computational linguistic tools
- Author
-
Eduardo Batista da Silva and Maurizio Babini
- Subjects
linguística de corpus ,terminologia ,redes neurais artificiais ,corpus linguistics ,terminology ,artificial neural networks ,Romanic languages ,PC1-5498 ,Philology. Linguistics ,P1-1091 - Abstract
O objetivo desse estudo é demonstrar, por meio de análise quantitativa e qualitativa, a eficácia de ferramentas linguístico-computacionais na seleção de terminologia para a produção de material terminológico. Serão apresentadas duas ferramentas linguístico-computacionais (WordSmith Tools e VocabProfile) e, também, sugestões para que o ensino de termos ofereça resultados práticos. A fundamentação teórico-metodológica recorreu a Barros (2004); Berber Sardinha (2000; 2005); Biderman (2001); Cabré (2007); Cobb (2007); Nation, (2003) e Sinclair (2004). O corpus da pesquisa foi constituído exclusivamente de material escrito na língua inglesa em diversas áreas de especialidade. Os procedimentos de preparação de material terminológico são exemplificados a partir de uma das áreas de especialidades utilizadas nos corpora de pesquisa, as Redes Neurais Artificiais. Os resultados obtidos indicam que a utilização do Wordsmith Tools juntamente com o VocabProfile pode fornecer dados importantes para a pesquisa linguistica.This paper aims to demonstrate by means of quantitative and qualitative analyses the effectiveness of the linguistic computational tools in selecting terminology for the production of terminological material. Two linguistic computational tools will be introduced (WordSmith Tools e VocabProfile) and also suggestions so as the teaching of terms may offer practical results. The theoretical-methodological approach relies on Barros (2004); Berber Sardinha (2000; 2005); Biderman (2001); Cabré (2007); Cobb (2007); Nation (2003) and Sinclair (2004). The research corpus was made solely of written material in English in several specialty languages. The procedures regarding terminological material preparation are exemplified with one of the specialty fields used in the research corpus, the Artificial Neural Networks. The obtained results indicate that the use of Wordsmith Tools in conjunction with VocabProfile might provide useful data for the linguistic research.
- Published
- 2011
- Full Text
- View/download PDF
42. Real Estate Evaluation engineering suported by multicriteria analysis and artificial neural network
- Author
-
MOREIRA, D. S., SILVA, R. S., and FERNANDES, A. M. R.
- Subjects
Multicriteria analysis ,Artificial Neural Networks ,TODIM ,Real Estate Market ,Electronic computers. Computer science ,QA75.5-76.95 - Abstract
To evaluate real estates means estimate its market value, taking into account several factors, such as: the particular characteristics of the property, market conditions and the balance between stakeholders in the negotiation. All these features make the evaluation a complex task, mainly because the real estate market represents an important segment of the national economy requiring more precise results. In this context, we conducted a survey on evaluation methodologies of properties that are used and identified that the main technique, regression analysis, has limitations that may compromise the outcome of the evaluation. This paper presents the research methodology developed using Multicriteria Analysis and Artificial Neural Network to evaluate residential apartments. The results were considered very satisfactory, indicating that the combination of multicriteria analysis and artificial neural network is efficient and promising.
- Published
- 2010
43. O Impacto da Taxa de Câmbio no Apreçamento de Opções no Brasil–Uma Análise Comparativa entre um Modelo de Rede Neural e o Modelo de Black & Scholes
- Author
-
Carlos Alberto Aragón de Planas, Léo da Rocha Ferreira, and Gerson Lachtermacher
- Subjects
options pricing ,artificial neural networks ,exchange rate ,Economic history and conditions ,HC10-1085 ,Economics as a science ,HB71-74 - Abstract
The main goal of this paper is to evaluate the impact of the exchange rate volatility in price prediction of derivative securities in the Brazilian capital markets using an artificial neural network technique, given the Black & Scholes Model limitations. For this purpose a multiplayer, feedforward neural network, trained by the backpropagation algorithm model, to perform the prediction of the Telemar option prices was developed. The model results show that price estimates are close to the real values, mainly when appended to the exchange rate, confirming that the performance of neural network is superior to other results. The inclusion of the exchange rate in neural networks technique results in a better price forecasting for the options, because the volatility in the price of the underlying asset is caused by temporary arbitrage of quotes among the national and foreign stock markets where the company is listed.
- Published
- 2009
44. Redes neurais artificiais na classificação de frutos: cenário bidimensional Fruit sorting using artificial neural networks: bidimensional case
- Author
-
Jean Paulo Silva Ramos
- Subjects
Redes Neurais Artificias ,classificação ,Redes Múltiplas Camadas ,Artificial Neural Networks ,back-propagation ,classifiers ,sorting ,Agriculture (General) ,S1-972 - Abstract
Inúmeras são as atividades agrícolas que necessitam de interação humana nos processos decisórios, e entre elas encontra-se a classificação de frutos. O consumo de frutos "in natura" exige altíssimo nível de qualidade, demandando um processo classificatório mais acurado. A classificação de frutos depende do reconhecimento de padrões natural ou artificial, de acordo com algumas categorias pré-definidas. Uma vez que um padrão de um fruto está sendo classificado, esse deve ser comparado com algum outro padrão armazenado. A maior parte da classificação de frutos é baseada na classificação humana.Este trabalho apresenta a possibilidade de uso de redes neurais artificiais no desenvolvimento de modelos de classificação de frutos por meio de vetores de padrões. Este trabalho foi desenvolvido no Departamento de Máquinas Agrícolas da Faculdade de Engenharia Agrícola da Universidade Estadual de Campinas, as redes neurais armazenaram os vetores de padrões de frutos peso, diâmetro. Esses componentes vetoriais associados entre si interagiram, determinando um vetor padrão de saída de acordo com os padrões de frutos armazenados. Para atingir esses objetivos, foi usada uma rede Perceptron de múltiplas camadas, com algoritmo de treinamento tipo retro-propagação para armazenar os vetores de padrões de frutos e para classificação desses padrões de entrada. A rede treinada conseguiu aprender a relação entre vetores de entrada e saída, demonstrando a potencialidade do uso de tais ferramentas na classificação artificial.Agriculture is one of the economic activities that more require the presence human being in the decision taking. Innumerable are the processes that require some type of human being interference in the conclusion of the processes. Fruit Sorting depends on human or artificial pattern recognition according to some pre defined categories. Once a fruit pattern is under classification, this one must be compared to some other ones stored. After that comparison it can be classified. Most sorting fruits jobs are human basis classification. This paper shows that using neural networks is possible to develop capable models of storing fruit pattern vectors. Given any fruit pattern vector to the model it can classify to the closest fruit pattern vector stored. The number of patterns were incremented and presented to the neural networks, classifying the presented fruits and proved the scalability of number of vector components used in fruit pattern vectors stored in the model. This work was developed in the Agricultural Machinery Department of the Agricultural Faculty in State University of Campinas, the neural networks stored fruit pattern vectors such as Weight, Diameter. These vectors components associated itself interacted determining an output pattern vector classifying according to the stored fruit vector patterns. A Multi Layer Perceptron Network with Backpropagation algorithm was used, storing the relationship between input fruit pattern vectors and output classification class vectors. The neural network was trained and tested presenting the desired results, it can be used as a tool for future fruit classification processes.
- Published
- 2003
- Full Text
- View/download PDF
45. Methodology for short-term horizon load forecasting using neural networks
- Author
-
Milke, Tafarel Franco, Abaide, Alzenira da Rosa, Santos, Laura Lisiane Callai dos, and Campos, Maurício de
- Subjects
Redes neurais artificiais ,Artificial neural networks ,Short term ,Curto prazo ,Previsão de carga ,ENGENHARIAS::ENGENHARIA ELETRICA [CNPQ] ,Load forecast - Abstract
Competitiveness and the insertion of new technologies in the electricity sector now condition companies to find ways to improve the quality of their services and ensure profitability. The short-term load forecasting activity is indispensable to support the planning and operation of electrical systems, aiming to make the energy supply stable and reliable. To perform load prediction using Artificial Neural Networks (ANN), it is necessary to evaluate the variables involved in the behavior of the daily load curve. By evaluating and obtaining the most available variables influencing the load behavior, it is then possible to use them as input to the adopted ANN model. Artificial neural networks are computational models inspired by the simplification of the functioning of biological neurons, with the ability to learn from experience with system inputs. They are similar to the brain due to the characteristics of knowledge acquired by a learning process and connections between its neurons used to store the acquired knowledge. A neural network has high power to generalize information after a learning phase, allowing to capture functional relationships between data producing output close to the expected. The process of learning or training the network consists in the application of ordered steps necessary for the tuning of the synaptic weights and thresholds of their neurons, aiming to produce the generalization of solutions by their outputs. The goal of network training is to make the application of a set of inputs a set of desired outputs. The tools using artificial intelligence techniques have been improved, allowing their application in various areas of knowledge, standing out among the main techniques used to perform short-term load forecasting, and are currently widely researched and employed for this purpose. Thus, its use has been showing more accurate results compared to traditional methods, since they can better develop the required mathematical processing. This paper presents a proposal for the prediction of the daily load curve for one day ahead applied to real energy, demand and temperature data, since it is the variables that best represent the short-term load behavior; For this, a model developed with multilayer perceptron neural networks using the Levenberg-Marquardt learning algorithm was implemented. The results found were satisfactory and acceptable compared to those presented in the literature review, being sufficient for practical application meeting the proposal of this work. Atualmente a competividade e a inserção de novas tecnologias no setor elétrico condicionam empresas a encontrar formas de melhorar a qualidade da prestação dos seus serviços e garantir lucratividade. A atividade de previsão de carga no curto prazo é indispensável para subsidiar o planejamento e a operação dos sistemas elétricos, visando tornar a oferta de energia estável e confiável. Para realizar a previsão de carga utilizando Redes Neurais Artificiais (RNA) é necessário avaliar as variáveis envolvidas no comportamento da curva de carga diária. Através da avaliação e obtenção das variáveis disponíveis mais influentes no comportamento da carga, é possível então utiliza-las como entrada do modelo RNA adotado. As redes neurais artificias são modelos computacionais inspirados na simplificação do funcionamento dos neurônios biológicos, com a capacidade de aprendizado a partir da experiência com as entradas do sistema. São semelhantes ao cérebro devido às características de conhecimento adquirido por um processo de aprendizagem e conexões entre seus neurônios utilizadas para armazenar o conhecimento adquirido. Uma rede neural possuiu alto poder de generalizar informações após uma fase de aprendizagem, possibilitando capturar relações funcionais entre os dados produzindo uma saída próxima daquela esperada. O processo de aprendizagem ou treinamento da rede consiste na aplicação de etapas ordenadas necessárias para que ocorra a sintonização dos pesos sinápticos e limiares de seus neurônios, visando à produção da generalização de soluções pelas suas saídas. O objetivo do treinamento da rede é tornar a aplicação de um conjunto de entradas em um conjunto de saídas desejadas. As ferramentas utilizando as técnicas de inteligência artificial vêm sendo aperfeiçoadas, permitindo a sua aplicação em diversas áreas do conhecimento, se destacando entre as principais técnicas utilizadas para realizar previsão de carga no curto prazo, sendo atualmente muito pesquisadas e empregadas para este fim. Desse modo, a sua utilização vem demonstrando resultados mais acurados em relação aos métodos tradicionais, pois conseguem desenvolver de melhor forma o processamento matemático requerido. Este trabalho apresenta uma proposta de previsão da curva de carga diária para um dia à frente aplicado a dados reais de energia, demanda e temperatura, pois são as variáveis que melhor representam o comportamento da carga no curto prazo; para isto foi implementado um modelo desenvolvido com redes neurais perceptron de múltiplas camadas, utilizando o algoritmo de aprendizagem Levenberg- Marquardt. Os resultados encontrados foram satisfatórios e aceitáveis comparados aos apresentados na revisão bibliográfica, sendo suficientes para aplicação prática atendendo a proposta deste trabalho.
- Published
- 2019
46. Redes neurais artificiais: Um método de representação comportamental
- Author
-
Monteiro, Érico Patrício
- Subjects
Artificial Intelligence ,Inteligência Artificial ,Representação do Conhecimento ,Knowledge Representation ,Artificial Neural Networks ,Redes Neurais Artificiais - Abstract
Anthropological researches, along the time, have pointed out that the main characteristic of human being is their thinking capacity, the solution procurement for their problems. This fact allied to the modern technologies represent, for the first time in history, that the knowledge keeps growing despite of controls, political differences, culture and region diversities. This paper represents a theoretical essay with the purpose of showing the evolution trends of carrying principles of human intelligence to the machines (Artificial Intelligence), whose main topic are the Artificial Neural Networks. The results of this study suggest that the transference evolution speed from human tasks to robots might be amplified, chiefly by Molecular Genetic and Quantum Physics researches. In contrast, the more this evolution grows, the more people be excluded from the process. Find outways to attenuate that seems to be the challenge of the millennium! Pesquisas antropológicas, ao longo do tempo, têm enfatizado ser a capacidade de pensar a principal responsável pela evolução dos seres humanos, de procurar soluções para seus problemas. Esse fato, aliado às modernas tecnologias, representa, pela primeira vez na história, que o conhecimento evolui a despeito de controles, diversidades políticas, culturais e de região. Esse artigo representa um ensaio teórico com o objetivo de mostrar as tendências de evolução da transmissão dos processos de inteligência humanos para as máquinas (Inteligência Artificial), cujo tópico principal são as Redes Neurais Artificiais. O resultado deste estudo sugere que a velocidade da transferência de tarefas humanas para robôs tende a ser ampliada, em função principalmente das pesquisas em Genética Molecular e Física Quântica. No entanto, quanto mais acelerado esse processo, maior a exclusão de grande parte da humanidade. Pesquisar caminhos que atenuem esse processo é o grande desafio do milênio!
- Published
- 2018
47. Development of technology based on artificial neural network for sign language gesture recognition
- Author
-
Silva, Brunna Carolinne Rocha, Calixto, Wesley Pacheco, Ramos, Carlos Fernando da Silva, Cruz Junior, Gelson da, Faria, Juliana Guimaraes, and Naka, Marco Hiroshi
- Subjects
Reconhecimento de padrões ,Redes neurais artificiais ,Tecnologia para reconhecimento ,Recognition of patterns ,Língua de sinais ,Artificial neural networks ,Technology for recognition ,Sign language ,CIENCIA DA COMPUTACAO [CIENCIAS EXATAS E DA TERRA] - Abstract
O intuito deste trabalho é projetar, desenvolver e avaliar quatro dispositivos capazes de identificar configuração, orientação e movimento das mãos, verificando qual possui melhor desempenho para reconhecimento de gestos da língua de sinais. A metodologia parte da definição do leiaute e dos componentes de aquisição e processamento de dados, da construção da base de dados tratados para cada gesto a ser reconhecido e da validação dos dispositivos propostos. São coletados sinais de sensores de flexão, acelerômetros e giroscópios, posicionados diferentemente em cada dispositivo. O reconhecimento dos padrões de cada gesto é realizado utilizando redes neurais artificiais. Após treinada, validada e testada, a rede neural interligada aos dispositivos obtêm média de acerto de até 96,8%. O dispositivo validado oferece eficácia e eficiência para identificar gestos da língua de sinais e demonstra que o uso da abordagem sensorial é promissora. The purpose of this paper is to design, develop and evaluate four devices capable of identifying configuration, orientation and movement of the hands, verifying which one has better performance recognition of sign language gestures. The methodology starts from the definition of the layout and the components of data acquisition and processing, the construction of the database treated for each gesture to be recognized and validation of the proposed devices. Signs of flex sensors, accelerometers and gyroscopes are collected, positioned differently on each device. The recognition of the patterns of each gesture is performed using artificial neural networks. After being trained, validated and tested, the neural network interconnected to the devices obtain a hit rate of up to 96.8%. The validated device offers efficacy and efficiency to identify sign language gestures and demonstrates that the use of the sensory approach is promising.
- Published
- 2018
48. Intelligent mushroom harvest prediction system proposal
- Author
-
Costa, Jorge, Branco, Frederico, Martins, José, Moreira, Fernando, Au-Yong-Oliveira, Manuel, Pérez-Cota, Manuel, González Castro, Miguel Ramón, and Díaz Rodríguez, María
- Subjects
Visão por computador ,Precision agriculture ,Artificial neural networks ,Processamento digital de imagem ,Redes neuronais artificiais ,Sistema de informação ,Computer vision ,Digital image processing ,Information system ,Agricultura de precisão - Abstract
As organizações do setor agroindustrial, atualmente têm apostado cada vez mais no desenvolvimento de sistemas tecnológicos, que permitem a informatização de todos os seus processos. Recentemente os métodos e técnicas de visão por computador têm sido muito utilizadas para monitorização e inspeção durante o período de produção e colheita, permitindo detetar problemas antecipadamente e com isto, melhorar a qualidade dos produtos. No setor da produção de cogumelos um dos aspetos mais importantes, e talvez mais preponderantes, é poder prever a sua produção. Com este propósito é proposto um Sistema Inteligente de Previsão de Colheita de Cogumelos (SIPCC), baseado em técnicas e métodos de visão por computador e Redes Neuronais Artificiais (RNA). Este trabalho expõe uma arquitetura de um SIPCC a nível funcional e técnica, complementada com a apresentação e análise de dados que demonstram a sua viabilidade. Organizations of the agro-industrial sector, are now increasingly investing in the development of technological systems that allow the computerization of all its processes. Recently the methods and techniques of computer vision have been widely used for monitoring and inspection during the production and harvesting, allowing detect problems early and thus, improve the quality of products. In the field of mushroom production one of the most important aspects, and perhaps most prevalent, is to be able to predict its production. To this end it is proposed an Intelligent System Mushroom Harvest Forecast (SIPCC), based on techniques and methods of computer vision and Artificial Neural Networks (ANN). This paper presents an architecture of a SIPCC functional and technical level, complemented with the analysis and presentation of data demonstrating its viability.
- Published
- 2018
49. Contribuição de parâmetros do entorno urbano sobre o ambiente térmico de um campus universitário
- Author
-
Debiazi, Pedro Renan and Souza, Léa Cristina Lucas de
- Subjects
Artificial neural networks ,urbano ,Geographical Information Systems ,Redesneuraisartificiais ,Sky view factor ,Urban climate ,Vegetação urbana ,Urban vegetation ,Fator de visão do céu ,Sistemas de informações geográficas - Abstract
Resumo Este artigo investiga a influência de algumas características do entorno urbano sobre a temperatura do ar, considerando parâmetros urbanos como o coeficiente de ocupação do solo (CO), coeficiente de vegetação urbana (CVU), fator de visão de céu (FVC) e coeficiente de cobertura do solo (CCS). O espaço amostral utilizado é o campus da Universidade Federal de São Carlos, em São Carlos, SP, Brasil. O método emprega dataloggers para medições de temperatura do arem diferentes pontos de coleta, além da determinação de índices urbanísticos nesses pontos amostrais. Os dados térmicos e urbanísticos são relacionados entre si por desenvolvimento de modelos deredes neurais artificiais (RNA), considerando-se três raios de abrangência ao redor dos pontos amostrais: 25 m, 50 m e 100 m. Dentre os modelos de previsão desenvolvidos, o de melhor desempenho é incorporado a um sistema de informação geográfica (SIG), permitindo a simulação de dados para outros pontos do campus e viabilizando a criação de mapas térmicos mais detalhados. Os resultados demonstram que o CVU é o elemento mais significativo na determinação do padrão térmico do campus e que os modelos de RNA associados à plataforma SIG podem ser instrumentos úteis para o apoio às ações que visem a qualidade térmica do campus. Abstract This paper investigates the influence of some characteristics of the urban environment on the air temperature by considering urban parameters such as occupancy coefficient (OC), urban vegetation coefficient (UVC), sky view factor (SVF) and cover coefficient (CC). The campus of the Federal University of São Carlos, in São Carlos, Brazil was used for spatial sampling. The method uses dataloggers for air temperature measurements in different collecting points, as well as the determination of urban indexes. The data of temperature and urban indexes are related to each other by the development of Artificial Neural Networks (ANN) models considering three radii around the sampling points: 25, 50 and 100 m. Among the prediction models developed, the one with the best performance is incorporated into a Geographical Information System (GIS), allowing data simulation of other points on the campus and leading to the creation of more detailed thermal maps. The results show that the UVC was the most significant element determining the thermal patterns in the campus. Furthermore, the ANN models associated with the GIS platform may be useful tools to support actions aiming at the thermal quality of the campus.
- Published
- 2017
50. Artificial neural networks for skin detection in digital images
- Author
-
Vicentini, Rafael Estéfano, Universidade Estadual Paulista (Unesp), and Lotufo, Anna Diva Plasencia [UNESP]
- Subjects
Redes neurais artificiais ,Reconhecimento de padrões ,Resilient backpropagation ,Artificial neural networks ,Pattern recognition ,Processamento digital de imagens ,Digital image processing - Abstract
Submitted by Rafael Estefano Vicentini null (rafaelvicentini@dee.feis.unesp.br) on 2017-12-14T18:01:32Z No. of bitstreams: 1 DISSERTAÇÃO-RAFAEL ESTÉFANO VICENTINI.pdf: 15039479 bytes, checksum: 43a2765c1d39e13b3435f194a64198ec (MD5) Approved for entry into archive by Cristina Alexandra de Godoy null (cristina@adm.feis.unesp.br) on 2017-12-18T10:48:16Z (GMT) No. of bitstreams: 1 vicentini_re_me_ilha.pdf: 15039479 bytes, checksum: 43a2765c1d39e13b3435f194a64198ec (MD5) Made available in DSpace on 2017-12-18T10:48:16Z (GMT). No. of bitstreams: 1 vicentini_re_me_ilha.pdf: 15039479 bytes, checksum: 43a2765c1d39e13b3435f194a64198ec (MD5) Previous issue date: 2017-10-20 Na última década, o aumento da capacidade de processamento de informação em computadores e dispositivos de uso pessoal possibilitou o desenvolvimento de filtros e classificadores automatizados que operam em tempo real, aplicados em diversas áreas. No âmbito do Processamento Digital de Imagens associado às Redes Neurais Artificiais, os filtros emulam a percepção humana buscando por padrões para identificação de características de interesse. Filtros que têm por objetivo restringir o acesso a conteúdo impróprio partem da identificação de pele - principal indício de presença humana em uma imagem. Independentemente de sua complexidade e/ou robustez, caso o classificador não seja capaz de identificar as diferentes tonalidades de pele sob diferentes condições de captura, sua eficácia é prejudicada. Frente à diversificada forma de descrever uma tonalidade de pele usando diferentes espaços de cor, neste estudo foram destacados os espaços de cor RGB, YCbCr e HSV, amplamente utilizados em equipamentos de captura (por exemplo câmeras fotográficas e filmadoras digitais). A partir de exemplos apresentados durante a etapa de treinamento, as RNAs devem estar aptas para classificar as tonalidades em dois grupos distintos: pele e não pele. Dentre os espaços de cores indicados, seja utilizando ou descartando o aspecto da iluminação (critério amplamente discutido na literatura), este trabalho busca avaliar qual possui a maior taxa de detecção de pele em uma imagem. Over the last decade, the increasing capacity of data processing in personal computers and devices could develop filters and automatic classifiers working in real time and applied in several areas. Considering Digital Image Processing and Artificial Neural Networks, these filters emulate the human perception searching for patterns to identify specific features. Filters which the main goal is to restrict the access to inappropriate content starts identifying skin tones - the main evidence of human presence in a picture. Although being complex and robust, if the classifier is not able to identify distinct skin tones under random capture conditions, the accuracy is minimal. Facing several ways on describing skin tones over different color spaces, this work uses the RGB, YCbCr and HSV color spaces which are widely applied in recording devices (photographic and digital cameras for example). Based on the examples shown during the training phase, the ANNs must be able to classify skin tones into two distinct groups: skin and non skin. Among the different color spaces used, considering or not the luminance aspect (widely discussed on papers), this work intends to evaluate which one has the highest detection accuracy to identify skin tone in such a picture.
- Published
- 2017
Discovery Service for Jio Institute Digital Library
For full access to our library's resources, please sign in.