Avaliação estatística de métodos de poda aplicados em neurônios intermediários da rede neural MLP

There are several papers on pruning methods in the artificial neural networks area. However, with rare exceptions, none of them presents an appropriate statistical evaluation of such methods. In this article, we proved statistically the ability of some methods to reduce the number of neurons of the hidden layer of a multilayer perceptron neural network (MLP), and to maintain the same landing of classification error of the initial net. They are evaluated seven pruning methods. The experimental investigation was accomplished on five groups of generated data and in two groups of real data. Three variables were accompanied in the study: apparent classification error rate in the test group (REA); number of hidden neurons, obtained after the application of the pruning method; and number of training/retraining epochs, to evaluate the computational effort. The non-parametric Friedman's test was used to do the statistical analysis.

Análise da incerteza na representação de classes de cobertura do solo urbano resultantes da aplicação de uma rede neural artificial

The great diversity of materials that characterizes the urban environment determines a structure of mixed classes in a classification of multiespectral images. In that sense, it is important to define an appropriate classification system using a non parametric classifier, that allows incorporating non spectral (such as texture) data to the process. They also allow analyzing the uncertainty associated to each class from the output alues of the network calculated in relation to each class. Considering these properties, an experiment was carried out. This experiment consisted in the application of an Artificial Neural Network aiming at the classification of the urban land cover of Presidente Prudente and the analysis of the uncertainty in the representation of the mapped thematic classes. The results showed that it is possible to discriminate the variations in the urban land cover through the application of an Artificial Neural Network. It was also possible to visualize the spatial variation of the uncertainty in the attribution of classes of urban land cover from the generated representations. The class characterized by a defined pattern as intermediary related to the impermeability of the urban soil presented larger ambiguity degree and, therefore, larger mixture.

Fontes de distorções da rede geodésica Brasileira

The Brazilian Geodetic Network started to be established in the early 40's, employing classical surveying methods, such as triangulation and trilateration. With the introduction of satellite positioning systems, such as TRANSIT and GPS, that network was densified. That data was adjusted by employing a variety of methods, yielding distortions in the network that need to be understood. In this work, we analyze and interpret study cases in an attempt to understand the distortions in the Brazilian network. For each case, we performed the network adjustment employing the GHOST software suite. The results show that the distortion is least sensitive to the removal of invar baselines in the classical network. The network would be more affected by the inexistence of Laplace stations and Doppler control points, with differences up to 4.5 m.

Análise da prescrição de medicamentos de pacientes hipertensos atendidos pelo SUS da rede municipal de saúde de Rincão - SP

The purpose of this study was to identify the drugs most often prescribed for hypertension at the Municipal Health Care Center of the town of Rincäo, State of São Paulo, Brazil, and the principal interactions arising from their association with other drugs, both anti-hypertensives and those in other classes. The study included 725 hypertensive patients registered at this health care center who were regularly seen by a physician every three months. Data were collected on age, sex, occurrence of diabetes, smoking, sedentary lifestyle and overweight, to obtain a profile of the hypertensive population of the area. Control records of all patients were available at the pharmacy in the health care center, where patients obtained their drugs once a month. Of the 725 patients, 38% were male and 62% female. Most (57%) were between 50 and 70 years of age, 21% used tobacco and 43% led a sedentary lifestyle. Single-drug therapy accounted for 33% of the prescriptions, multidrug therapy for 66%. In addition to anti-hypertensives, 50% of the patients took drugs of other therapeutic classes. Of those receiving multidrug therapy, 34% used three or more anti-hypertensives and 66% used only two of these drugs. Drug interactions were detected in as many as 47% of the prescriptions. Captopril was the drug that showed most interactions with others (54%), followed by hydrochlorothiazide (27%), furosemide (14%), propanolol (4%), and nifedipine (1%). The analysis revealed that drug consumption by the patients investigated is high, with a concomitantly high number of episodes of drug interaction.

Estimação de frequência usando sensores de erro com atrasos adaptativos

This paper proposes a solution to improve the performance of the first order Early Error Sensing (EES) Adaptive Time Delay Tanlock Loops (ATDTL) presented in (Al-Zaabi, Al-Qutayri e Al-Araji, 2005), regarding to frequency estimation and tracking time. The EES-ATDTL are phaselocked-loops (PLL) used to hardware implementations, due to their simple structure. Fixed-points theorems are used to determine conditions for rapid convergence of the estimation process and a estimative of the frecuency input is obtained with a Gaussian filter that improves the gain adaptation. The mathematical models are the presented by (Al-Araji, Al-Qutayri e Al-Zaabi, 2006). Simulations have been performed to evaluate the theoretical results.

Processamento da ração no desempenho de juvenis de jundiá (Rhamdia voulezi) cultivados em tanques-rede

The aim of this study was to evaluate the zootechnical performance and chemical composition of jundia juveniles Rhandia voulezi fed with mash, pelletized and extruded diets and cultivated in cages. One hundred and fifty fish were distributed in fifteen cages (0.20 m3), arranged in cages (5m3), in the hydroelectric plant of Governador José Richa, in Iguaçu River (Paraná-Brazil). The fish were fed five times a day, with mash, pelletized and extruded diets, containing 30% of protein and 3250 Kcal/kg of digestible energy/kg. At the end of 60 days, significant differences (p<0.05) were observed on fish performance. The fish that were fed with pelletized diet, presented higher average of final weight (44.08g), final length (16.51cm) and weight gain (36.65g). The different processes did not influence fish survival. The averages of protein and mineral matter of carcass chemical composition did not differ statistically (p<0.05), while the values of humidity and lipideos presented differences among treatments. The highest average of lipideos (8.59%) was observed in peletized diet. The mash diet provided fish with lower lipideo (5.45%) and higher humidity (75.98%) rates. A pelleted diet is most suitable for juvenile catfish R. voulezi cultivation in cages.

Uso de sig e fotografias aéreas na determinação de conflitos na rede de drenagem na bacia experimental do Rio Pardo

This study aimed to detect and evaluate potential conflicts in the land use along the Rio Pardo's tributaries and sub-tributaries from its source which is in Pardinho, SP to the water capture system by the Basic Sanitation Company located in Botucatu, SP. It was used 2005 colored aerial photos, taking as a tool for assessing the Geographical Information System - IDRISI Andes. The results showed that the riparian forests were substituted predominantly by pastures, which is called conflict. In other words, the 1590.82ha of environmental preservation areas (PPA's) that should be protected by riparian forests and meadows (41.51%) in 2005, were used as pastures, being in conflict with the legislation. The aerial pictures associated to SIG-Idrisi proved to be efficient tools in mapping the area and monitoring the environmental impacts.

Volumes de tanques-rede na produção da tilápia-do-nilo: estudo de caso

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Avaliação do desempenho de tilápias (Oreochromis niloticus, Trewavas, 1983) em tanques-rede na represa de Jurumirim/Alto Rio Paranapanema

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Monitoramento de características físico-químicas da água e taxas de sedimentação de tripton no Córrego da Arribada (Baixo Tietê-SP) relacionadas à piscicultura em tanques-rede

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Efeito da densidade de estocagem na sobrevivência e crescimento do pampo (Trachinotus carolinus) em tanques-rede marinhos

Pós-graduação em Aquicultura - FCAV

Desenvolvimento ovariano do Prochilodus lineatus em dois sistemas de produção: tanques rede e viveiros escavados

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

DGPS em Rede: desenvolvimento e implantação via internet utilizando a Rede GNSS do Estado de São Paulo

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Análise da incerteza na representação de classes temáticas resultantes da aplicação de uma rede neural artificial

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Integridade, disponibilidade e acurácia no posicionamento RTK e RTK em rede: Investigação no contexto da rede GNSS ativa do Estado de São Paulo

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)