21 resultados para Gradient Descent method
em Reposit
Resumo:
A calibration method was developed using flow injection analysis (FI) with a Gradient Calibration Method (GCM). The method allows the rapid determination of zinc In foods (approximately 30 min) after treatment with concentrated sulphuric acid and 30% hydrogen peroxide, and analysis with flame atomic absorption spectrometry (FAAS). The method provides analytical results with a relative standard deviation of about 2% and requires less time than by conventional FI calibration. The electronic selection of different segments along the gradient and monitoring of the technique covers wide concentration ranges while maintaining the inherent high precision of flow injection analysis. Concentrations, flow rates, and flow times of the reagents were optimized in order to obtain best accuracy and precision. Flow rates of 10 mL/min were selected for zinc. In addition, the system enables electronic dilution and calibration where a multipoint curve can be constructed using a single sample injection.
Resumo:
Discriminative training of Gaussian Mixture Models (GMMs) for speech or speaker recognition purposes is usually based on the gradient descent method, in which the iteration step-size, ε, uses to be defined experimentally. In this letter, we derive an equation to adaptively determine ε, by showing that the second-order Newton-Raphson iterative method to find roots of equations is equivalent to the gradient descent algorithm. © 2010 IEEE.
Resumo:
Continuous-time neural networks for solving convex nonlinear unconstrained;programming problems without using gradient information of the objective function are proposed and analyzed. Thus, the proposed networks are nonderivative optimizers. First, networks for optimizing objective functions of one variable are discussed. Then, an existing one-dimensional optimizer is analyzed, and a new line search optimizer is proposed. It is shown that the proposed optimizer network is robust in the sense that it has disturbance rejection property. The network can be implemented easily in hardware using standard circuit elements. The one-dimensional net is used as a building block in multidimensional networks for optimizing objective functions of several variables. The multidimensional nets implement a continuous version of the coordinate descent method.
Resumo:
This work presents a procedure for electric load forecasting based on adaptive multilayer feedforward neural networks trained by the Backpropagation algorithm. The neural network architecture is formulated by two parameters, the scaling and translation of the postsynaptic functions at each node, and the use of the gradient-descendent method for the adjustment in an iterative way. Besides, the neural network also uses an adaptive process based on fuzzy logic to adjust the network training rate. This methodology provides an efficient modification of the neural network that results in faster convergence and more precise results, in comparison to the conventional formulation Backpropagation algorithm. The adapting of the training rate is effectuated using the information of the global error and global error variation. After finishing the training, the neural network is capable to forecast the electric load of 24 hours ahead. To illustrate the proposed methodology it is used data from a Brazilian Electric Company. © 2003 IEEE.
Resumo:
Pós-graduação em Matemática - IBILCE
Resumo:
Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)
Resumo:
The objective of this study was to evaluate the quality of bovine frozen-thawed sperm cells after Percoll gradient centrifugation. Frozen semen doses were obtained from six bulls of different breeds, including three taurine and three Zebu animals. Four ejaculates per bull were evaluated before and after discontinuous Percoll gradient centrifugation. Sperm motility was assessed by computer-assisted semen analysis and the integrity of the plasma and acrosomal membranes, as well as mitochondrial function, were evaluated using a combination of fluorescent probes propidium iodide, fluorescein isothiocyanate-conjugated Pisum sativum agglutinin and 5,5',6,6'-tetrachloro-1,1',3,3'-tetraethylbenzimidazolcarbocyanine iodide. The procedure of Percoll gradient centrifugation increased the percentage of total and progressive sperm motility, beat frequency, rectilinear motility, linearity and rapidly moving cells. In addition, the percentage of cells with intact plasma membrane and mitochondrial membrane potential was increased in post-centrifugation samples. However, the percentage of sperm cells with intact acrosomal membrane was markedly reduced. The method used selected the motile cells with intact plasma membrane and higher mitochondrial functionality in frozen-thawed bull semen, but processing, centrifugation and/or the Percoll medium caused damage to the acrosomal membrane.
Resumo:
Propolis obtained from honeybee hives has been widely used in medicine, cosmetics, and industry due to its versatile biological activities (antioxidant, antimicrobial, fungicidal, antiviral, antiulcer, immunostimulating, and cytostatic). These activities are mainly attributed to the presence of flavonoids in propolis, which points out the interest in quantifying these constituents in propolis preparations, as well as validation of analytical methodologies. High-performance liquid chromatography (HPLC) methods have been reported to quantify isolated flavonoids or these compounds in complex biological matrices, such as herbal raw materials and extractive preparations. An efficient, precise, and reliable method was developed for quantification of propolis extractive solution using HPLC with UV detection. The chromatograms were obtained from various gradient elution systems (GES) tested in order to establish the ideal conditions for the analysis of propolis extractive solution, using methanol and water: acetonitrile (97.5 : 2.5, v/v) as mobile phase. Gradient reversed phase chromatography was performed using a stainless steel column (250 x 4.6 mm i.d., 5 mum) filled with Chromsep RP 18 (Varian), column temperature at 30.0 +/- 0.1degreesC and detection at 310 nm. The main validation parameters of the method were also determined. The method showed linearity for chrysin in the range 0.24-2.4 mug mL(-1) with good correlation coefficients (0.9975). Precision and accuracy were determined. The obtained results demonstrate the efficiency of the proposed method. The analytical procedure is reliable and offers advantages in terms of speed and cost of reagents.
Resumo:
The usual design of field-emission displays (FEDs) often results in high vacuum pressure gradients inside the glass plates of the device, and this is the main limitation to the widespread availability of large area FEDs. In this paper, we perform theoretical calculations using the finite element method for determining the pressure distributions in several pumping configurations, including a new FED configuration known as porous FED or pFED. The approach here is capable of clarifying the design issues influencing the final pressure relevant to the field-emission display devices. (C) 2001 Elsevier B.V. B.V. All rights reserved.
Resumo:
A reversed phase liquid chromatographic method was developed for the simultaneous determination of carboxylic acids and phenolics in white wines. The samples, diluted, were injected onto a Spherisorb ODS-2 column with a gradient of sulfuric acid (pH 2.5)/methanol as mobile phase. A diode array detector was used which was set at 210nm for carboxylic acids and altered to 278nm, during the run, far phenolics and sorbic acid. The identification of compounds was based on retention time, co-chromatography and UV spectrum. Some clean-up methods (sep-pak C-18 and an ion exchange column) mere tested and did not improve the results.The analysis was simple, with no sample preparation. Application of this method was illustrated by analyses of Brazilian Welchriesling wines.
Resumo:
Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)
Resumo:
Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)
Resumo:
An analytical approach for the spin stabilized satellite attitude propagation is presented using the non-singular canonical variables to describe the rotational motion. Two sets of variables were introduced for Fukushima in 1994 by a canonical transformation and they are useful when the angle between z-satellite axis of a coordinate system fixed in artificial satellite and the rotational angular momentum vector is zero or when the angle between Z-equatorial axis and rotation angular momentum vector is zero. Analytical solutions for rotational motion equations and torque-free motion are discussed in terms of the elliptic functions and by the application of some simplification to get an approximated solution. These solutions are compared with a numerical solution and the results show a good agreement for many rotation periods. When the mean Hamiltonian associated with the gravity gradient torque is included, an analytical solution is obtained by the application of the successive approximations' method for the satellite in an elliptical orbit. These solutions show that the magnitude of the rotation angular moment is not affected by the gravity gradient torque but this torque causes linear and periodic variations in the angular variables, long and short periodic variations in Z-equatorial component of the rotation angular moment and short periodic variations in x-satellite component of the rotation angular moment. The goal of this analysis is to emphasize the geometrical and physical meaning of the non-singular variables and to validate the approximated analytical solution for the rotational motion without elliptic functions for a non-symmetrical satellite. The analysis can be applied for spin stabilized satellite and in this case the general solution and the approximated solution are coincidence. Then the results can be used in analysis of the space mission of the Brazilian Satellites. (C) 2007 COSPAR. Published by Elsevier Ltd. All rights reserved.
Resumo:
This paper presents vectorized methods of construction and descent of quadtrees that can be easily adapted to message passing parallel computing. A time complexity analysis for the present approach is also discussed. The proposed method of tree construction requires a hash table to index nodes of a linear quadtree in the breadth-first order. The hash is performed in two steps: an internal hash to index child nodes and an external hash to index nodes in the same level (depth). The quadtree descent is performed by considering each level as a vector segment of a linear quadtree, so that nodes of the same level can be processed concurrently. © 2012 Springer-Verlag.
Resumo:
Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)