Approximation of hyperbolic tangent activation function using hybrid methods

Artificial Neural Networks are widely used in various applications in engineering, as such solutions of nonlinear problems. The implementation of this technique in reconfigurable devices is a great challenge to researchers by several factors, such as floating point precision, nonlinear activation function, performance and area used in FPGA. The contribution of this work is the approximation of a nonlinear function used in ANN, the popular hyperbolic tangent activation function. The system architecture is composed of several scenarios that provide a tradeoff of performance, precision and area used in FPGA. The results are compared in different scenarios and with current literature on error analysis, area and system performance. © 2013 IEEE.

As contribuições da ciência cognitiva à composição musical

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Modelo do bulbo olfativo baseado em redes neurais recorrentes

Pós-graduação em Física - IGCE

Risco de crédito: uma abordagem utilizando análise discriminante, regressão logística e redes neurais artificiais

Considering the relevance of researches concerning credit risk, model diversity and the existent indicators, this thesis aimed at verifying if the Fleuriet Model contributes in discriminating Brazilian open capital companies in the analysis of credit concession. We specifically intended to i) identify the economic-financial indicators used in credit risk models; ii) identify which economic-financial indicators best discriminate companies in the analysis of credit concession; iii) assess which techniques used (discriminant analysis, logistic regression and neural networks) present the best accuracy to predict company bankruptcy. To do this, the theoretical background approached the concepts of financial analysis, which introduced themes relative to the company evaluation process; considerations on credit, risk and analysis; Fleuriet Model and its indicators, and, finally, presented the techniques for credit analysis based on discriminant analysis, logistic regression and artificial neural networks. Methodologically, the research was defined as quantitative, regarding its nature, and explanatory, regarding its type. It was developed using data derived from bibliographic and document analysis. The financial demonstrations were collected by means of the Economática ® and the BM$FBOVESPA website. The sample was comprised of 121 companies, being those 70 solvents and 51 insolvents from various sectors. In the analyses, we used 22 indicators of the Traditional Model and 13 of the Fleuriet Model, totalizing 35 indicators. The economic-financial indicators which were a part of, at least, one of the three final models were: X1 (Working Capital over Assets), X3 (NCG over Assets), X4 (NCG over Net Revenue), X8 (Type of Financial Structure), X9 (Net Thermometer), X16 (Net Equity divided by the total demandable), X17 (Asset Turnover), X20 (Net Equity Profitability), X25 (Net Margin), X28 (Debt Composition) and X31 (Net Equity over Asset). The final models presented setting values of: 90.9% (discriminant analysis); 90.9% (logistic regression) and 97.8% (neural networks). The modeling in neural networks presented higher accuracy, which was confirmed by the ROC curve. In conclusion, the indicators of the Fleuriet Model presented relevant results for the research of credit risk, especially if modeled by neural networks.

An assessment of the economic impact of climate change on the water sector in the Turks and Caicos Islands

The best description of water resources for Grand Turk was offered by Pérez Monteagudo (2000) who suggested that rain water was insufficient to ensure a regular water supply although water catchment was being practised and water catchment possibilities had been analysed. Limestone islands, mostly flat and low lying, have few possibilities for large scale surface storage, and groundwater lenses exist in very delicate equilibrium with saline seawater, and are highly likely to collapse due to sea level rise, improper extraction, drought, tidal waves or other extreme event. A study on the impact of climate change on water resources in the Turks and Caicos Islands is a challenging task, due to the fact that the territory of the Islands covers different environmental resources and conditions, and accurate data are lacking. The present report is based on collected data wherever possible, including grey data from several sources such as the Intergovernmental Panel on Climate Change (IPCC) and Cuban meteorological service data sets. Other data were also used, including the author’s own estimates and modelling results. Although challenging, this was perhaps the best approach towards analysing the situation. Furthermore, IPCC A2 and B2 scenarios were used in the present study in an effort to reduce uncertainty. The main conclusion from the scenario approach is that the trend observed in precipitation during the period 1961 - 1990 is decreasing. Similar behaviour was observed in the Caribbean region. This trend is associated with meteorological causes, particularly with the influence of the North Atlantic Anticyclone. The annual decrease in precipitation is estimated to be between 30-40% with uncertain impacts on marine resources. After an assessment of fresh water resources in Turks and Caicos Islands, the next step was to estimate residential water demand based on a high fertility rate scenario for the Islands (one selected from four scenarios and compared to countries having similar characteristics). The selected scenario presents higher projections on consumption growth, enabling better preparation for growing water demand. Water demand by tourists (stopover and excursionists, mainly cruise passengers) was also obtained, based on international daily consumption estimates. Tourism demand forecasts for Turks and Caicos Islands encompass the forty years between 2011 and 2050 and were obtained by means of an Artificial Neural Networks approach. for the A2 and B2 scenarios, resulting in the relation BAU>B2>A2 in terms of tourist arrivals and water demand levels from tourism. Adaptation options and policies were analysed. Resolving the issue of the best technology to be used for Turks and Caicos Islands is not directly related to climate change. Total estimated water storage capacity is about 1, 270, 800 m3/ year with 80% capacity load for three plants. However, almost 11 desalination plants have been detected on Turks and Caicos Islands. Without more data, it is not possible to estimate long term investment to match possible water demand and more complex adaptation options. One climate change adaptation option would be the construction of elevated (30 metres or higher) storm resistant water reservoirs. The unit cost of the storage capacity is the sum of capital costs and operational and maintenance costs. Electricity costs to pump water are optional as water should, and could, be stored for several months. The costs arising for water storage are in the range of US$ 0.22 cents/m3 without electricity costs. Pérez Monteagudo (2000) estimated water prices at around US$ 2.64/m3 in stand points, US$ 7.92 /m3 for government offices, and US$ 13.2 /m3for cistern truck vehicles. These data need to be updated. As Turks and Caicos Islands continues to depend on tourism and Reverse Osmosis (RO) for obtaining fresh water, an unavoidable condition to maintaining and increasing gross domestic product(GDP) and population welfare, dependence on fossil fuels and vulnerability to increasingly volatile prices will constitute an important restriction. In this sense, mitigation supposes a synergy with adaptation. Energy demand and emissions of carbon dioxide (CO2) were also estimated using an emissions factor of 2. 6 tCO2/ tonne of oil equivalent (toe). Assuming a population of 33,000 inhabitants, primary energy demand was estimated for Turks and Caicos Islands at 110,000 toe with electricity demand of around 110 GWh. The business as usual (BAU), as well as the mitigation scenarios were estimated. The BAU scenario suggests that energy use should be supported by imported fossil fuels with important improvements in energy efficiency. The mitigation scenario explores the use of photovoltaic and concentrating solar power, and wind energy. As this is a preliminary study, the local potential and locations need to be identified to provide more relevant estimates. Macroeconomic assumptions are the same for both scenarios. By 2050, Turks and Caicos Islands could demand 60 m toe less than for the BAU scenario.

Utilização de redes neurais artificiais no ajuste de controladores suplementares e dispositivo FACTS STATCOM para a melhoria da estabilidade a pequenas perturbações do sistema elétrico de potência

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Modelos de regressão aleatória, análise multivariada e redes neurais artificiais na avaliação genética da produção de leite de vacas Holandesas

Pós-graduação em Zootecnia - FCAV

Aplicação da rede neural MLP (Multilayer Perceptron) em indústria de pisos e revestimentos do Pólo Cerâmico de Santa Gertrudes - SP

Pós-graduação em Geociências e Meio Ambiente - IGCE

Construct validity of behavioral models of anxiety: where experimental psychopathology meets ecology and evolution

In experimental psychopathology, construct validity is usually enhanced by addressing theories from other fields in its nomological network. In the field of anxiety research, this construct is related to antipredator behavior, conserved across phylogeny in its functions and neural basis, but not necessarily on its topography. Even though the relations between behavioral models of anxiety and statements from behavioral ecology and evolutionary biology are commonly made in anxiety research, these are rarely tested, at least explicitly. However, in order to increase construct validity in experimental anxiety, testing predictions from those theories is highly desirable. This article discusses these questions, suggesting a few ways in which behavioral ecological and evolutionary hypotheses of anxiety-like behavior may be tested.

Modelos para previsão de carga a curto prazo através de redes neurais artificiais com treinamento baseado na teoria da informação

O conhecimento prévio do valor da carga é de extrema importância para o planejamento e operação dos sistemas de energia elétrica. Este trabalho apresenta os resultados de um estudo investigativo da aplicação de Redes Neurais Artificiais do tipo Perceptron Multicamadas com treinamento baseado na Teoria da Informação para o problema de Previsão de Carga a curto prazo. A aprendizagem baseada na Teoria da Informação se concentra na utilização da quantidade de informação (Entropia) para treinamento de uma rede neural artificial. Dois modelos previsores são apresentados sendo que os mesmos foram desenvolvidos a partir de dados reais fornecidos por uma concessionária de energia. Para comparação e verificação da eficiência dos modelos propostos um terceiro modelo foi também desenvolvido utilizando uma rede neural com treinamento baseado no critério clássico do erro médio quadrático. Os resultados alcançados mostraram a eficiência dos sistemas propostos, que obtiveram melhores resultados de previsão quando comparados ao sistema de previsão baseado na rede treinada pelo critério do MSE e aos sistemas previsores já apresentados na literatura.

Detecção de refletores sísmicos por rede neural discreta

As redes neurais artificiais têm provado serem uma poderosa técnica na resolução de uma grande variedade de problemas de otimização. Nesta dissertação é desenvolvida uma nova rede neural, tipo recorrente, sem realimentação (self-feedback loops) e sem neurônios ocultos, para o processamento do sinal sísmico, para fornecer a posição temporal, a polaridade e as amplitudes estimadas dos refletores sísmicos, representadas pelos seus coeficientes de reflexão. A principal característica dessa nova rede neural consiste no tipo de função de ativação utilizada, a qual permite três possíveis estados para o neurônio. Busca-se estimar a posição dos refletores sísmicos e reproduzir as verdadeiras polaridades desses refletores. A idéia básica desse novo tipo de rede, aqui denominada rede neural discreta (RND), é relacionar uma função objeto, que descreve o problema geofísico, com a função de Liapunov, que descreve a dinâmica da rede neural. Deste modo, a dinâmica da rede leva a uma minimização local da sua função de Liapunov e consequentemente leva a uma minimização da função objeto. Assim, com uma codificação conveniente do sinal de saída da rede tem-se uma solução do problema geofísico. A avaliação operacional da arquitetura desta rede neural artificial é realizada em dados sintéticos gerados através do modelo convolucional simples e da teoria do raio. A razão é para explicar o comportamento da rede com dados contaminados por ruído, e diante de pulsos fonte de fases mínima, máxima e misturada.

Disturbance detection for optimal database storage in electrical distribution systems using artificial immune systems with negative selection

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Predicting soluble solid content in intact jaboticaba [Myrciaria jaboticaba (Vell.) O. Berg] fruit using near-infrared spectroscopy and chemometrics

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Pruning methods to MLP neural networks considering proportional apparent error rate for classification problems with unbalanced data

This article deals with classification problems involving unequal probabilities in each class and discusses metrics to systems that use multilayer perceptrons neural networks (MLP) for the task of classifying new patterns. In addition we propose three new pruning methods that were compared to other seven existing methods in the literature for MLP networks. All pruning algorithms presented in this paper have been modified by the authors to do pruning of neurons, in order to produce fully connected MLP networks but being small in its intermediary layer. Experiments were carried out involving the E. coli unbalanced classification problem and ten pruning methods. The proposed methods had obtained good results, actually, better results than another pruning methods previously defined at the MLP neural network area. (C) 2014 Elsevier Ltd. All rights reserved.

Preliminary diagnosis of ophtalmological diseases through machine learning techniques

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)