Máquina de vetores-suporte intervalar

The Support Vector Machines (SVM) has attracted increasing attention in machine learning area, particularly on classification and patterns recognition. However, in some cases it is not easy to determinate accurately the class which given pattern belongs. This thesis involves the construction of a intervalar pattern classifier using SVM in association with intervalar theory, in order to model the separation of a pattern set between distinct classes with precision, aiming to obtain an optimized separation capable to treat imprecisions contained in the initial data and generated during the computational processing. The SVM is a linear machine. In order to allow it to solve real-world problems (usually nonlinear problems), it is necessary to treat the pattern set, know as input set, transforming from nonlinear nature to linear problem. The kernel machines are responsible to do this mapping. To create the intervalar extension of SVM, both for linear and nonlinear problems, it was necessary define intervalar kernel and the Mercer s theorem (which caracterize a kernel function) to intervalar function

Contribuições aos Processos de Clustering com Base em Métricas não-Euclidianas

In this work we present a new clustering method that groups up points of a data set in classes. The method is based in a algorithm to link auxiliary clusters that are obtained using traditional vector quantization techniques. It is described some approaches during the development of the work that are based in measures of distances or dissimilarities (divergence) between the auxiliary clusters. This new method uses only two a priori information, the number of auxiliary clusters Na and a threshold distance dt that will be used to decide about the linkage or not of the auxiliary clusters. The number os classes could be automatically found by the method, that do it based in the chosen threshold distance dt, or it is given as additional information to help in the choice of the correct threshold. Some analysis are made and the results are compared with traditional clustering methods. In this work different dissimilarities metrics are analyzed and a new one is proposed based on the concept of negentropy. Besides grouping points of a set in classes, it is proposed a method to statistical modeling the classes aiming to obtain a expression to the probability of a point to belong to one of the classes. Experiments with several values of Na e dt are made in tests sets and the results are analyzed aiming to study the robustness of the method and to consider heuristics to the choice of the correct threshold. During this work it is explored the aspects of information theory applied to the calculation of the divergences. It will be explored specifically the different measures of information and divergence using the Rényi entropy. The results using the different metrics are compared and commented. The work also has appendix where are exposed real applications using the proposed method

Técnicas neuronais de agrupamentos aplicadas à detecção de anomalias em mamogramas digitais

This work proposes the development of an intelligent system for analysis of digital mammograms, capable to detect and to classify masses and microcalcifications. The digital mammograms will be pre-processed through techniques of digital processing of images with the purpose of adapting the image to the detection system and automatic classification of the existent calcifications in the suckles. The model adopted for the detection and classification of the mammograms uses the neural network of Kohonen by the algorithm Self Organization Map - SOM. The algorithm of Vector quantization, Kmeans it is also used with the same purpose of the SOM. An analysis of the performance of the two algorithms in the automatic classification of digital mammograms is developed. The developed system will aid the radiologist in the diagnosis and accompaniment of the development of abnormalities

Sistema auxiliar na análise de mamogramas digitais

This work proposes the development of a Computer System for Analysis of Mammograms SCAM, that aids the doctor specialist in the identification and analysis of existent lesions in digital mammograms. The computer system for digital mammograms processing will make use of a group of techniques of Digital Image Processing (DIP), with the purpose of aiding the medical professional to extract the information contained in the mammogram. This system possesses an interface of easy use for the user, allowing, starting from the supplied mammogram, a group of processing operations, such as, the enrich of the images through filtering techniques, the segmentation of areas of the mammogram, the calculation the area of the lesions, thresholding the lesion, and other important tools for the medical professional's diagnosis. The Wavelet Transform will used and integrated into the computer system, with the objective of allowing a multiresolution analysis, thus supplying a method for identifying and analyzing microcalcifications

Algoritmo SOM com estrutura hierárquica e dinâmica aplicado a compressão de imagens

ln this work the implementation of the SOM (Self Organizing Maps) algorithm or Kohonen neural network is presented in the form of hierarchical structures, applied to the compression of images. The main objective of this approach is to develop an Hierarchical SOM algorithm with static structure and another one with dynamic structure to generate codebooks (books of codes) in the process of the image Vector Quantization (VQ), reducing the time of processing and obtaining a good rate of compression of images with a minimum degradation of the quality in relation to the original image. Both self-organizing neural networks developed here, were denominated HSOM, for static case, and DHSOM, for the dynamic case. ln the first form, the hierarchical structure is previously defined and in the later this structure grows in an automatic way in agreement with heuristic rules that explore the data of the training group without use of external parameters. For the network, the heuristic mIes determine the dynamics of growth, the pruning of ramifications criteria, the flexibility and the size of children maps. The LBO (Linde-Buzo-Oray) algorithm or K-means, one ofthe more used algorithms to develop codebook for Vector Quantization, was used together with the algorithm of Kohonen in its basic form, that is, not hierarchical, as a reference to compare the performance of the algorithms here proposed. A performance analysis between the two hierarchical structures is also accomplished in this work. The efficiency of the proposed processing is verified by the reduction in the complexity computational compared to the traditional algorithms, as well as, through the quantitative analysis of the images reconstructed in function of the parameters: (PSNR) peak signal-to-noise ratio and (MSE) medium squared error

Aplicação de técnicas de aprendizado de máquina no reconhecimento de classes estruturais de proteínas

Nowadays, classifying proteins in structural classes, which concerns the inference of patterns in their 3D conformation, is one of the most important open problems in Molecular Biology. The main reason for this is that the function of a protein is intrinsically related to its spatial conformation. However, such conformations are very difficult to be obtained experimentally in laboratory. Thus, this problem has drawn the attention of many researchers in Bioinformatics. Considering the great difference between the number of protein sequences already known and the number of three-dimensional structures determined experimentally, the demand of automated techniques for structural classification of proteins is very high. In this context, computational tools, especially Machine Learning (ML) techniques, have become essential to deal with this problem. In this work, ML techniques are used in the recognition of protein structural classes: Decision Trees, k-Nearest Neighbor, Naive Bayes, Support Vector Machine and Neural Networks. These methods have been chosen because they represent different paradigms of learning and have been widely used in the Bioinfornmatics literature. Aiming to obtain an improvment in the performance of these techniques (individual classifiers), homogeneous (Bagging and Boosting) and heterogeneous (Voting, Stacking and StackingC) multiclassification systems are used. Moreover, since the protein database used in this work presents the problem of imbalanced classes, artificial techniques for class balance (Undersampling Random, Tomek Links, CNN, NCL and OSS) are used to minimize such a problem. In order to evaluate the ML methods, a cross-validation procedure is applied, where the accuracy of the classifiers is measured using the mean of classification error rate, on independent test sets. These means are compared, two by two, by the hypothesis test aiming to evaluate if there is, statistically, a significant difference between them. With respect to the results obtained with the individual classifiers, Support Vector Machine presented the best accuracy. In terms of the multi-classification systems (homogeneous and heterogeneous), they showed, in general, a superior or similar performance when compared to the one achieved by the individual classifiers used - especially Boosting with Decision Tree and the StackingC with Linear Regression as meta classifier. The Voting method, despite of its simplicity, has shown to be adequate for solving the problem presented in this work. The techniques for class balance, on the other hand, have not produced a significant improvement in the global classification error. Nevertheless, the use of such techniques did improve the classification error for the minority class. In this context, the NCL technique has shown to be more appropriated

Compressão auto-adaptativa de imagens coloridas

Image compress consists in represent by small amount of data, without loss a visual quality. Data compression is important when large images are used, for example satellite image. Full color digital images typically use 24 bits to specify the color of each pixel of the Images with 8 bits for each of the primary components, red, green and blue (RGB). Compress an image with three or more bands (multispectral) is fundamental to reduce the transmission time, process time and record time. Because many applications need images, that compression image data is important: medical image, satellite image, sensor etc. In this work a new compression color images method is proposed. This method is based in measure of information of each band. This technique is called by Self-Adaptive Compression (S.A.C.) and each band of image is compressed with a different threshold, for preserve information with better result. SAC do a large compression in large redundancy bands, that is, lower information and soft compression to bands with bigger amount of information. Two image transforms are used in this technique: Discrete Cosine Transform (DCT) and Principal Component Analysis (PCA). Primary step is convert data to new bands without relationship, with PCA. Later Apply DCT in each band. Data Loss is doing when a threshold discarding any coefficients. This threshold is calculated with two elements: PCA result and a parameter user. Parameters user define a compression tax. The system produce three different thresholds, one to each band of image, that is proportional of amount information. For image reconstruction is realized DCT and PCA inverse. SAC was compared with JPEG (Joint Photographic Experts Group) standard and YIQ compression and better results are obtain, in MSE (Mean Square Root). Tests shown that SAC has better quality in hard compressions. With two advantages: (a) like is adaptive is sensible to image type, that is, presents good results to divers images kinds (synthetic, landscapes, people etc., and, (b) it need only one parameters user, that is, just letter human intervention is required

Desenvolvimento de um reator fotoquímico aplicável no tratamento de efluentes fenólicos presentes na indústria do petróleo

With water pollution increment at the last years, so many progresses in researches about treatment of contaminated waters have been developed. In wastewaters containing highly toxic organic compounds, which the biological treatment cannot be applied, the Advanced Oxidation Processes (AOP) is an alternative for degradation of nonbiodegradable and toxic organic substances, because theses processes are generation of hydroxyl radical based on, a highly reactivate substance, with ability to degradate practically all classes of organic compounds. In general, the AOP request use of special ultraviolet (UV) lamps into the reactors. These lamps present a high electric power demand, consisting one of the largest problems for the application of these processes in industrial scale. This work involves the development of a new photochemistry reactor composed of 12 low cost black light fluorescent lamps (SYLVANIA, black light, 40 W) as UV radiation source. The studied process was the photo-Fenton system, a combination of ferrous ions, hydrogen peroxide, and UV radiation, it has been employed for the degradation of a synthetic wastewater containing phenol as pollutant model, one of the main pollutants in the petroleum industry. Preliminary experiments were carrier on to estimate operational conditions of the reactor, besides the effects of the intensity of radiation source and lamp distribution into the reactor. Samples were collected during the experiments and analyzed for determining to dissolved organic carbon (DOC) content, using a TOC analyzer Shimadzu 5000A. The High Performance Liquid Chromatography (HPLC) was also used for identification of the cathecol and hydroquinone formed during the degradation process of the phenol. The actinometry indicated 9,06⋅1018 foton⋅s-1 of photons flow, for 12 actived lamps. A factorial experimental design was elaborated which it was possible to evaluate the influence of the reactants concentration (Fe2+ and H2O2) and to determine the most favorable experimental conditions ([Fe2+] = 1,6 mM and [H2O2] = 150,5 mM). It was verified the increase of ferrous ions concentration is favorable to process until reaching a limit when the increase of ferrous ions presents a negative effect. The H2O2 exhibited a positive effect, however, in high concentrations, reaching a maximum ratio degradation. The mathematical modeling of the process was accomplished using the artificial neural network technique

Terra Natal, de Ferreira Itajubá, e a permanência do romantismo

In this dissertation we argue that the work Terra Natal, by the poet Ferreira Itajubá, presents, predominantly, themes and forms typical of the Romantic period, although it came out in 1914, when the romantic aesthetic was no longer popular in Brazilian Literature. Firstly, we discuss about Ferreira Itajubá s biography, his bibliographic production as well as the critical fortune of his works. Secondly, we outline the course of the Romantic Movement from its emergence in Europe to its arrival in Brazil as well as its impact in the state of Rio Grande do Norte. Finally, we analyze of Terra Natal, taking into account other aspects such as context and structure through the analysis of the poems VII, IX, XVIII, XXI and also of the song Viola , inserted in the poem VIII. We highlight the following romantic themes during the analysis: exile, nostalgia for the past, cult of native land, the romantic view of woman, projection of soul s state on nature, and religiosity. We will try to create dialogues between Terra Natal and works of popular Romantic Brazilian writers, such as Gonçalves Dias, Casimiro de Abreu, José de Alencar,Visconde de Taunay and Castro Alves, in order to show the affiliation of the above-referenced work and the Romantic ideals

Modernismo e tradição da oralidade na poesia :uma leitura de Clã do jabuti, de Mário de Andrade, e Catimbó, de Ascenso Ferreira

Estudio sobre los libros Clã do jabuti, de Mário de Andrade, y Catimbó, de Ascenso Ferreira, publicados en el año de 1927. Dentro del contexto del Modernismo en los años de 1920, se evidencia la filiación de los poetas con las tendencias literarias modernista y regionalista. Eso se confirma por los elementos introducidos en sus poesías, en las referidas obras, que traen el sentido de brasilidade discutido en la época, específicamente en el momento post1924. Tales elementos se refieren a las marcas de la oralidad que fueron aprovechadas esteticamente en los referidos textos, en lo que toca a la representación que poseen en el ámbito de la tradición de la cultura popular brasileña. En ese sentido, fue establecida una relación entre la moderna literatura nacional y la tradición de la cultura popular. Los textos observados muestran la diversidad cultural del país por medio de tales marcas, distinguiéndose por el aspecto nacionalista de Mário de Andrade, en consonancia con el proyecto de nación idealizado, y por el aspecto nacionalista/regionalista de Ascenso Ferreira, saliendo al encuentro de elementos nacionales y regionales. Siendo así, Clã do jabuti es un repertorio de todo Brasil y representación de la cultura nacional, de acuerdo con la brasilidade propuesta a partir de 1924; y Catimbó es la cristalización poética, en tonos de una brasilidade más norteña, de las particularidades de la región

Construções com tópico marcado em cartas pessoais brasileiras dos séculos XVIII, XIX E XX

This dissertation is a research on the marked topic construction (CT) in Brazilian personal letters from eighteenth, nineteenth and twentieth centuries. The goal of our research is to verify if CT are present in the writing of Brazilians born in the centuries in question. Our research focus is based on the assumptions of generative theory (CHOMSKY 1981; 1986), which states that grammar is internalized in the mind / brain of the writers, with the emphasis on studies of grammatical change, as pointed texts by Paixão de Sousa (2004), Carneiro (2005); Galves, Namiuti and Paixão de Sousa (2006) and Martins (2009). Our corpus was extracted from Projeto Para a História do Português Brasileiro (PHPB) and Cartas Brasileiras coletânea de fontes para o estudo do português. We selected forty-six correspondents who should be inserted into the two criteria set out in this research: to be Brazilian and be born in the centuries mentioned above, so that we could find legitimate topic constructions of PB. This work is based on researches by Pontes (1987), Mateus et al. (2003), Araujo (2006, 2009), Berlinck, Duarte and Oliveira (2009), which actively support us in the study of this linguistic phenomenon in Portuguese. The results show that the marked topic construction in our corpus appear on the writing of Brazilians since the second half of the eighteenth century, while the typical constructions in Brazilian Portuguese locative topic, subject topic and copy topic - are already reflected in the I-language of the writers born in the second half of the nineteenth century and the first of the twentieth century

Floração precoce em cana-de-açúcar - um estudo utilizando ferramentas de análise in silico e proteômica

Sugarcane is one of the most important products of the world and Brazil is responsible for 25 % of the world production. One problem of this culture at northeast of Brazil is the early flowering. In our laboratory, it has been made before four subtractive libraries using early and late flowering genotypes in order to identify messages related to the flowering process. In this work, two cDNAs were chosen to make in silico analysis and overexpression constructs. Another approach to understand the flowering process in sugarcane was to use proteomic tools. First, the protocol for protein extraction using apical meristem was set up. After that, these proteins were separated on two bidimensional gels. It was possible to observe some difference for some regions of these gels as well as some proteins that can be found in all conditions. The next step, spots will be isolated and sequence on MS spectrometry in order to understand this physiological process in sugarcane

Otimização em comitês de classificadores: uma abordagem baseada em filtro para seleção de subconjuntos de atributos

Traditional applications of feature selection in areas such as data mining, machine learning and pattern recognition aim to improve the accuracy and to reduce the computational cost of the model. It is done through the removal of redundant, irrelevant or noisy data, finding a representative subset of data that reduces its dimensionality without loss of performance. With the development of research in ensemble of classifiers and the verification that this type of model has better performance than the individual models, if the base classifiers are diverse, comes a new field of application to the research of feature selection. In this new field, it is desired to find diverse subsets of features for the construction of base classifiers for the ensemble systems. This work proposes an approach that maximizes the diversity of the ensembles by selecting subsets of features using a model independent of the learning algorithm and with low computational cost. This is done using bio-inspired metaheuristics with evaluation filter-based criteria

Uma análise da aplicação do modelo de Rede Neural RePART em Comitês de classificadores

RePART (Reward/Punishment ART) is a neural model that constitutes a variation of the Fuzzy Artmap model. This network was proposed in order to minimize the inherent problems in the Artmap-based model, such as the proliferation of categories and misclassification. RePART makes use of additional mechanisms, such as an instance counting parameter, a reward/punishment process and a variable vigilance parameter. The instance counting parameter, for instance, aims to minimize the misclassification problem, which is a consequence of the sensitivity to the noises, frequently presents in Artmap-based models. On the other hand, the use of the variable vigilance parameter tries to smoouth out the category proliferation problem, which is inherent of Artmap-based models, decreasing the complexity of the net. RePART was originally proposed in order to minimize the aforementioned problems and it was shown to have better performance (higer accuracy and lower complexity) than Artmap-based models. This work proposes an investigation of the performance of the RePART model in classifier ensembles. Different sizes, learning strategies and structures will be used in this investigation. As a result of this investigation, it is aimed to define the main advantages and drawbacks of this model, when used as a component in classifier ensembles. This can provide a broader foundation for the use of RePART in other pattern recognition applications

Um estudo algorítmico de problemas logísticos na indústria de petróleo e gás natural

This work consists on the study of two important problems arising from the operations of petroleum and natural gas industries. The first problem the pipe dimensioning problem on constrained gas distribution networks consists in finding the least cost combination of diameters from a discrete set of commercially available ones for the pipes of a given gas network, such that it respects minimum pressure requirements at each demand node and upstream pipe conditions. On its turn, the second problem the piston pump unit routing problem comes from the need of defining the piston pump unit routes for visiting a number of non-emergent wells in on-shore fields, i.e., wells which don t have enough pressure to make the oil emerge to surface. The periodic version of this problem takes into account the wells re-filling equation to provide a more accurate planning in the long term. Besides the mathematical formulation of both problems, an exact algorithm and a taboo search were developed for the solution of the first problem and a theoretical limit and a ProtoGene transgenetic algorithm were developed for the solution of the second problem. The main concepts of the metaheuristics are presented along with the details of their application to the cited problems. The obtained results for both applications are promising when compared to theoretical limits and alternate solutions, either relative to the quality of the solutions or to associated running time