Análise de Agrupamentos Com Base na Teoria da Informação: Uma Abordagem Representativa

Currently, one of the biggest challenges for the field of data mining is to perform cluster analysis on complex data. Several techniques have been proposed but, in general, they can only achieve good results within specific areas providing no consensus of what would be the best way to group this kind of data. In general, these techniques fail due to non-realistic assumptions about the true probability distribution of the data. Based on this, this thesis proposes a new measure based on Cross Information Potential that uses representative points of the dataset and statistics extracted directly from data to measure the interaction between groups. The proposed approach allows us to use all advantages of this information-theoretic descriptor and solves the limitations imposed on it by its own nature. From this, two cost functions and three algorithms have been proposed to perform cluster analysis. As the use of Information Theory captures the relationship between different patterns, regardless of assumptions about the nature of this relationship, the proposed approach was able to achieve a better performance than the main algorithms in literature. These results apply to the context of synthetic data designed to test the algorithms in specific situations and to real data extracted from problems of different fields

Não-extensividade em percolação por ligações de longo - alcance

A linear chain do not present phase transition at any finite temperature in a one dimensional system considering only first neighbors interaction. An example is the Ising ferromagnet in which his critical temperature lies at zero degree. Analogously, in percolation like disordered geometrical systems, the critical point is given by the critical probability equals to one. However, this situation can be drastically changed if we consider long-range bonds, replacing the probability distribution by a function like . In this kind of distribution the limit α → ∞ corresponds to the usual first neighbor bond case. In the other hand α = 0 corresponds to the well know "molecular field" situation. In this thesis we studied the behavior of Pc as a function of a to the bond percolation specially in d = 1. Our goal was to check a conjecture proposed by Tsallis in the context of his Generalized Statistics (a generalization to the Boltzmann-Gibbs statistics). By this conjecture, the scaling laws that depend with the size of the system N, vary in fact with the quantitie

Segmentação Fuzzy de Texturas e Vídeos

The segmentation of an image aims to subdivide it into constituent regions or objects that have some relevant semantic content. This subdivision can also be applied to videos. However, in these cases, the objects appear in various frames that compose the videos. The task of segmenting an image becomes more complex when they are composed of objects that are defined by textural features, where the color information alone is not a good descriptor of the image. Fuzzy Segmentation is a region-growing segmentation algorithm that uses affinity functions in order to assign to each element in an image a grade of membership for each object (between 0 and 1). This work presents a modification of the Fuzzy Segmentation algorithm, for the purpose of improving the temporal and spatial complexity. The algorithm was adapted to segmenting color videos, treating them as 3D volume. In order to perform segmentation in videos, conventional color model or a hybrid model obtained by a method for choosing the best channels were used. The Fuzzy Segmentation algorithm was also applied to texture segmentation by using adaptive affinity functions defined for each object texture. Two types of affinity functions were used, one defined using the normal (or Gaussian) probability distribution and the other using the Skew Divergence. This latter, a Kullback-Leibler Divergence variation, is a measure of the difference between two probability distributions. Finally, the algorithm was tested in somes videos and also in texture mosaic images composed by images of the Brodatz album

Dinâmica e estrutura de redes complexas no modelo de afinidade

In this work we elaborate and discuss a Complex Network model which presents connectivity scale free probability distribution (power-law degree distribution). In order to do that, we modify the rule of the preferential attachment of the Bianconi-Barabasi model, including a factor which represents the similarity of the sites. The term that corresponds to this similarity is called the affinity, and is obtained by the modulus of the difference between the fitness (or quality) of the sites. This variation in the preferential attachment generates very interesting results, by instance the time evolution of the connectivity, which follows a power-law distribution ki / ( t t0 )fi, where fi indicates the rate to the site gain connections. Certainly this depends on the affinity with other sites. Besides, we will show by numerical simulations results for the average path length and for the clustering coefficient

Universaidade em sistemas da mecâniaestatística de não equilíbrio com estados absorventes e percolação geográfica

Complex systems have stimulated much interest in the scientific community in the last twenty years. Examples this area are the Domany-Kinzel cellular automaton and Contact Process that are studied in the first chapter this tesis. We determine the critical behavior of these systems using the spontaneous-search method and short-time dynamics (STD). Ours results confirm that the DKCA e CP belong to universality class of Directed Percolation. In the second chapter, we study the particle difusion in two models of stochastic sandpiles. We characterize the difusion through diffusion constant D, definite through in the relation h(x)2i = 2Dt. The results of our simulations, using finite size scalling and STD, show that the diffusion constant can be used to study critical properties. Both models belong to universality class of Conserved Directed Percolation. We also study that the mean-square particle displacement in time, and characterize its dependence on the initial configuration and particle density. In the third chapter, we introduce a computacional model, called Geographic Percolation, to study watersheds, fractals with aplications in various areas of science. In this model, sites of a network are assigned values between 0 and 1 following a given probability distribution, we order this values, keeping always its localization, and search pk site that percolate network. Once we find this site, we remove it from the network, and search for the next that has the network to percole newly. We repeat these steps until the complete occupation of the network. We study the model in 2 and 3 dimension, and compare the bidimensional case with networks form at start real data (Alps e Himalayas)

The Wigner function associated with the Rogers-Szego polynomials

A Wigner function associated with the Rogers-Szego polynomials is proposed and its properties are discussed. It is shown that from such a Wigner function it is possible to obtain well-behaved probability distribution functions for both angle and action variables, defined on the compact support -pi less than or equal to theta < pi, and for m greater than or equal to 0, respectively. The width of the angle probability density is governed by the free parameter q characterizing the polynomials.

Testing option pricing with the Edgeworth expansion

There is a well-developed framework, the Black-Scholes theory, for the pricing of contracts based on the future prices of certain assets, called options. This theory assumes that the probability distribution of the returns of the underlying asset is a Gaussian distribution. However, it is observed in the market that this hypothesis is flawed, leading to the introduction of a fudge factor, the so-called volatility smile. Therefore, it would be interesting to explore extensions of the Black-Scholes theory to non-Gaussian distributions. In this paper, we provide an explicit formula for the price of an option when the distributions of the returns of the underlying asset is parametrized by an Edgeworth expansion, which allows for the introduction of higher independent moments of the probability distribution, namely skewness and kurtosis. We test our formula with options in the Brazilian and American markets, showing that the volatility smile can be reduced. We also check whether our approach leads to more efficient hedging strategies of these instruments. (C) 2004 Elsevier B.V. All rights reserved.

Microscopic origin of non-Gaussian distributions of financial returns

In this paper we study the possible microscopic origin of heavy-tailed probability density distributions for the price variation of financial instruments. We extend the standard log-normal process to include another random component in the so-called stochastic volatility models. We study these models under an assumption, akin to the Born-Oppenheimer approximation, in which the volatility has already relaxed to its equilibrium distribution and acts as a background to the evolution of the price process. In this approximation, we show that all models of stochastic volatility should exhibit a scaling relation in the time lag of zero-drift modified log-returns. We verify that the Dow-Jones Industrial Average index indeed follows this scaling. We then focus on two popular stochastic volatility models, the Heston and Hull-White models. In particular, we show that in the Hull-White model the resulting probability distribution of log-returns in this approximation corresponds to the Tsallis (t-Student) distribution. The Tsallis parameters are given in terms of the microscopic stochastic volatility model. Finally, we show that the log-returns for 30 years Dow Jones index data is well fitted by a Tsallis distribution, obtaining the relevant parameters. (c) 2007 Elsevier B.V. All rights reserved.

Determination and statistical treatment of short-circuit currents for faults on the lines of networks of 440 kV CESP system

The distribution of short-circuit current is investigated by means of two methods, one direct and the other analytic; both methods consider uniform probability distribution of line faults. In the direct method, the procedure consists of calculating fault currents at equidistant points along the line, starting from one of the end points and considering the other end open. The magnitude of the current is classified according to Brazilian standards (regulation NBR-7118). The analytic method assumes that the distribution of short-circuit currents through the busbar and the distribution of the line length connected to it are known, as well as the independence of values. The method is designed to determine the probability that fault currents through a line will surpass the pre-established magnitude, thus generating frequency distribution curves of short-circuit currents along the lines.

Segmentação de dados de perfilamento a laser em áreas urbanas utilizando uma abordagem Bayesiana

In this paper is presented a region-based methodology for Digital Elevation Model segmentation obtained from laser scanning data. The methodology is based on two sequential techniques, i.e., a recursive splitting technique using the quad tree structure followed by a region merging technique using the Markov Random Field model. The recursive splitting technique starts splitting the Digital Elevation Model into homogeneous regions. However, due to slight height differences in the Digital Elevation Model, region fragmentation can be relatively high. In order to minimize the fragmentation, a region merging technique based on the Markov Random Field model is applied to the previously segmented data. The resulting regions are firstly structured by using the so-called Region Adjacency Graph. Each node of the Region Adjacency Graph represents a region of the Digital Elevation Model segmented and two nodes have connectivity between them if corresponding regions share a common boundary. Next it is assumed that the random variable related to each node, follows the Markov Random Field model. This hypothesis allows the derivation of the posteriori probability distribution function whose solution is obtained by the Maximum a Posteriori estimation. Regions presenting high probability of similarity are merged. Experiments carried out with laser scanning data showed that the methodology allows to separate the objects in the Digital Elevation Model with a low amount of fragmentation.

Fluxo de potência trifásico probabilístico para redes de distribuição usando o método de estimação por pontos

In this paper the point estimation method is applied to solve the probabilistic power flow problem for unbalanced three-phase distribution systems. Through the implementation of this method the probability distribution functions of voltages (magnitude and angle) as well as the active and reactive power flows in the branches of the distribution system are determined. Two different approaches of the point estimation method are presented (2m and 2m+1 point schemes). In order to test the proposed methodology, the IEEE 34 and 123 bus test systems are used. The results obtained with both schemes are compared with the ones obtained by a Monte Carlo Simulation (MCS).

Solução da equação de Fokker-Planck para o potencial do oscilador harmônico duplo

Pós-graduação em Biofísica Molecular - IBILCE

Uma abordagem probabilística do número de reprodução básica em modelos epidemiológicos com aplicação na ferrugem do eucalipto

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Citricultura na região sul paulista: modelo de previsão econômica para tomada de decisão em condições de risco

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Análise energética e econômica da cultura de algodão em sistemas agrícolas familiares

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)