Estudo Comparativo de Métricas de Pontuação para Aprendizagem Estrutural de Redes Bayesianas

Bayesian networks are powerful tools as they represent probability distributions as graphs. They work with uncertainties of real systems. Since last decade there is a special interest in learning network structures from data. However learning the best network structure is a NP-Hard problem, so many heuristics algorithms to generate network structures from data were created. Many of these algorithms use score metrics to generate the network model. This thesis compare three of most used score metrics. The K-2 algorithm and two pattern benchmarks, ASIA and ALARM, were used to carry out the comparison. Results show that score metrics with hyperparameters that strength the tendency to select simpler network structures are better than score metrics with weaker tendency to select simpler network structures for both metrics (Heckerman-Geiger and modified MDL). Heckerman-Geiger Bayesian score metric works better than MDL with large datasets and MDL works better than Heckerman-Geiger with small datasets. The modified MDL gives similar results to Heckerman-Geiger for large datasets and close results to MDL for small datasets with stronger tendency to select simpler network structures

Modelagem de Superfícies Seletivas de Freqüência e Antenas de Microfita utilizando Redes Neurais Artificiais

This work has as main objective the application of Artificial Neural Networks, ANN, in the resolution of problems of RF /microwaves devices, as for example the prediction of the frequency response of some structures in an interest region. Artificial Neural Networks, are presently a alternative to the current methods of analysis of microwaves structures. Therefore they are capable to learn, and the more important to generalize the acquired knowledge, from any type of available data, keeping the precision of the original technique and adding the low computational cost of the neural models. For this reason, artificial neural networks are being increasily used for modeling microwaves devices. Multilayer Perceptron and Radial Base Functions models are used in this work. The advantages/disadvantages of these models and the referring algorithms of training of each one are described. Microwave planar devices, as Frequency Selective Surfaces and microstrip antennas, are in evidence due the increasing necessities of filtering and separation of eletromagnetic waves and the miniaturization of RF devices. Therefore, it is of fundamental importance the study of the structural parameters of these devices in a fast and accurate way. The presented results, show to the capacities of the neural techniques for modeling both Frequency Selective Surfaces and antennas

Sistema não-Intrusivo para Estimação da Direção do olhar utilizando redes neurais artificiais

The fundamental senses of the human body are: vision, hearing, touch, taste and smell. These senses are the functions that provide our relationship with the environment. The vision serves as a sensory receptor responsible for obtaining information from the outside world that will be sent to the brain. The gaze reflects its attention, intention and interest. Therefore, the estimation of gaze direction, using computer tools, provides a promising alternative to improve the capacity of human-computer interaction, mainly with respect to those people who suffer from motor deficiencies. Thus, the objective of this work is to present a non-intrusive system that basically uses a personal computer and a low cost webcam, combined with the use of digital image processing techniques, Wavelets transforms and pattern recognition, such as artificial neural network models, resulting in a complete system that performs since the image acquisition (including face detection and eye tracking) to the estimation of gaze direction. The obtained results show the feasibility of the proposed system, as well as several feature advantages.

Verificação formal automatizada para sistemas de raciocínio procedural (PRS) utilizando redes de petri coloridas (RPC)

Este trabalho apresenta uma técnica de verificação formal de Sistemas de Raciocínio Procedural, PRS (Procedural Reasoning System), uma linguagem de programação que utiliza a abordagem do raciocínio procedural. Esta técnica baseia-se na utilização de regras de conversão entre programas PRS e Redes de Petri Coloridas (RPC). Para isso, são apresentadas regras de conversão de um sub-conjunto bem expressivo da maioria da sintaxe utilizada na linguagem PRS para RPC. A fim de proceder fia verificação formal do programa PRS especificado, uma vez que se disponha da rede de Petri equivalente ao programa PRS, utilizamos o formalismo das RPCs (verificação das propriedades estruturais e comportamentais) para analisarmos formalmente o programa PRS equivalente. Utilizamos uma ferramenta computacional disponível para desenhar, simular e analisar as redes de Petri coloridas geradas. Uma vez que disponhamos das regras de conversão PRS-RPC, podemos ser levados a querer fazer esta conversão de maneira estritamente manual. No entanto, a probabilidade de introdução de erros na conversão é grande, fazendo com que o esforço necessário para garantirmos a corretude da conversão manual seja da mesma ordem de grandeza que a eliminação de eventuais erros diretamente no programa PRS original. Assim, a conversão automatizada é de suma importância para evitar que a conversão manual nos leve a erros indesejáveis, podendo invalidar todo o processo de conversão. A principal contribuição deste trabalho de pesquisa diz respeito ao desenvolvimento de uma técnica de verificação formal automatizada que consiste basicamente em duas etapas distintas, embora inter-relacionadas. A primeira fase diz respeito fias regras de conversão de PRS para RPC. A segunda fase é concernente ao desenvolvimento de um conversor para fazer a transformação de maneira automatizada dos programas PRS para as RPCs. A conversão automática é possível, porque todas as regras de conversão apresentadas seguem leis de formação genéricas, passíveis de serem incluídas em algoritmos

Predição não-linear de curvas de produção de petróleo via redes neurais recursivas

One of the main activities in the petroleum engineering is to estimate the oil production in the existing oil reserves. The calculation of these reserves is crucial to determine the economical feasibility of your explotation. Currently, the petroleum industry is facing problems to analyze production due to the exponentially increasing amount of data provided by the production facilities. Conventional reservoir modeling techniques like numerical reservoir simulation and visualization were well developed and are available. This work proposes intelligent methods, like artificial neural networks, to predict the oil production and compare the results with the ones obtained by the numerical simulation, method quite a lot used in the practice to realization of the oil production prediction behavior. The artificial neural networks will be used due your learning, adaptation and interpolation capabilities

Desenvolvimento de sistemas embarcados para redes de sensores e atuadores sem fio aplicadas em unidades de elevação de petróleo do tipo Plunger-Lift

The Wireless Sensor Networks (WSN) methods applied to the lifting of oil present as an area with growing demand technical and scientific in view of the optimizations that can be carried forward with existing processes. This dissertation has as main objective to present the development of embedded systems dedicated to a wireless sensor network based on IEEE 802.15.4, which applies the ZigBee protocol, between sensors, actuators and the PLC (Programmable Logic Controller), aiming to solve the present problems in the deployment and maintenance of the physical communication of current elevation oil units based on the method Plunger-Lift. Embedded systems developed for this application will be responsible for acquiring information from sensors and control actuators of the devices present at the well, and also, using the Modbus protocol to make this network becomes transparent to the PLC responsible for controlling the production and delivery information for supervisory SISAL

Um sistema para a inferência de variáveis usando redes neurais artificiais

This work aims to obtain a low-cost virtual sensor to estimate the quality of LPG. For the acquisition of data from a distillation tower, software HYSYS ® was used to simulate chemical processes. These data will be used for training and validation of an Artificial Neural Network (ANN). This network will aim to estimate from available simulated variables such as temperature, pressure and discharge flow of a distillation tower, the mole fraction of pentane present in LPG. Thus, allowing a better control of product quality

e-Resíduos: A influência da norma européia WEEE na estratégia da indústria de celulares no Brasil e no mundo e o impacto ambiental do descarte inadequado

This present Thesis, is explorer work and presents an analysis of e-wastes of the industry of cellular mobile telephony, evaluating the evolution of the telecommunications nets and as if it holds the global and Brazilian market of cellular telephony. It approaches the elements gifts in the cellular devices that can badly cause to the environment and the health, the discarding of the devices in end of life cycle is made. It analyzes the new European regulation of electric equipment residues and electronic, the WEEE, as it influenced the strategy of the companies manufacturers of mobile phone cellular and of that she forms is possible to create a Brazilian national industry for recycling of devices of cellular, with conditions to globally competition. For this some possible models of being implanted in Brazil are presented. The project of law 203/91 on solid residues is argued and as it would be interesting if to persist some proposals presented to the project, to create a Brazilian market of recycling with capacity of global competition for use to advantage of the European regulation if to get a competitive advantage

Uma proposta para preparação de pequenas e médias empresas brasileiras prestadoras de serviços ao setor de telecomunicações para implantação da metodologia Seis Sigma

The objective of this work consists of considering a script so that operating companies in the telecommunications sector, mainly small and medium-sized enterprises, can reach excellency in its operations and get, thus, conditions to compete with companies consolidated in its niche of market. The proposal intends to lead the enterprises to a level of qualification of processes that to become them apt to adopt the Six Sigma method as part of its culture. From the analysis of the essential processes of the sector, methods and tools will be suggested for guarantee the continuous improvement of these processes, without forgetting the internal peculiarities of each company

Análise de desempenho de vendas em telecomunicações utilizando técnicas de mineração de dados

Nowadays, telecommunications is one of the most dynamic and strategic areas in the world. Organizations are always seeking to find new management practices within an ever increasing competitive environment where resources are getting scarce. In this scenario, data obtained from business and corporate processes have even greater importance, although this data is not yet adequately explored. Knowledge Discovery in Databases (KDD) appears then, as an option to allow the study of complex problems in different areas of management. This work proposes both a systematization of KDD activities using concepts from different methodologies, such as CRISP-DM, SEMMA and FAYYAD approaches and a study concerning the viability of multivariate regression analysis models to explain corporative telecommunications sales using performance indicators. Thus, statistical methods were outlined to analyze the effects of such indicators on the behavior of business productivity. According to business and standard statistical analysis, equations were defined and fit to their respective determination coefficients. Tests of hypotheses were also conducted on parameters with the purpose of validating the regression models. The results show that there is a relationship between these development indicators and the amount of sales

Avaliação de economias de escala e escopo no desempenho de equipes de serviços aplicando Análise de Envoltória de Dados DEA: um estudo em Empresa de Telecomunicações no Brasil

This Master s Thesis proposes the application of Data Envelopment Analysis DEA to evaluate economies of scale and economies of scope in the performance of service teams involved with installation of data communication circuits, based on the study of a major telecommunication company in Brazil. Data was collected from the company s Operational Performance Division. Initial analysis of a data set, including nineteen installation teams, was performed considering input oriented methods. Subsequently, the need for restrictions on weights is analyzed using the Assurance Region method, checking for the existence of zero-valued weights. The resulting returns to scale are then verified. Further analyses using the Assurance Region Constant (AR-I-C) and Variable (AR-I-V) models verify the existence of variable, rather than constant, returns to scale. Therefore, all of the final comparisons use scores obtained through the AR-I-V model. In sequence, we verify if the system has economies of scope by analyzing the behavior of the scores in terms of individual or multiple outputs. Finally, conventional results, used by the company in study to evaluate team performance, are compared to those generated using the DEA methodology. The results presented here show that DEA is a useful methodology for assessing team performance and that it may contribute to improvements on the quality of the goal setting procedure.

Análise de retorno de escala e tamanho do mercado no desempenho de equipes de vendas aplicando análise de envoltória de dados DEA: um estudo em empresa de Telecomunicações no Brasil

This Master s Thesis proposes the application of Data Envelopment Analysis DEA to evaluate the performance of sales teams, based on a study of their coverage areas. Data was collected from the company contracted to distribute the products in the state of Ceará. Analyses of thirteen sales coverage areas were performed considering first the output-oriented constant return to scale method (CCR-O), then this method with assurance region (AR-O-C) and finally the method of variable returns to scale with assurance region (AR-O-V). The method used in the first approach is shown to be inappropriate for this study, since it inconveniently generates zero-valued weights, allowing that an area under evaluation obtain the maximal score by not producing. Using weight restrictions, through the assurance region methods AR-O-C and AR-O-V, decreasing returns to scale are identified, meaning that the improvement in performance is not proportional to the size of the areas being analyzed. Observing data generated by the analysis, a study is carried out, aiming to design improvement goals for the inefficient areas. Complementing this study, GDP data for each area was compared with scores obtained using AR-O-V analysis. The results presented in this work show that DEA is a useful methodology for assessing sales team performance and that it may contribute to improvements on the quality of the management process.

Busca heurística através de algoritmo genético e memético com construção de vocábulos para o problema de atribuição de localidades a anéis Sonet

Telecommunications play a key role in contemporary society. However, as new technologies are put into the market, it also grows the demanding for new products and services that depend on the offered infrastructure, making the problems of planning telecommunications networks, despite the advances in technology, increasingly larger and complex. However, many of these problems can be formulated as models of combinatorial optimization, and the use of heuristic algorithms can help solving these issues in the planning phase. In this project it was developed two pure metaheuristic implementations Genetic algorithm (GA) and Memetic Algorithm (MA) plus a third hybrid implementation Memetic Algorithm with Vocabulary Building (MA+VB) for a problem in telecommunications that is known in the literature as Problem SONET Ring Assignment Problem or SRAP. The SRAP arises during the planning stage of the physical network and it consists in the selection of connections between a number of locations (customers) in order to meet a series of restrictions on the lowest possible cost. This problem is NP-hard, so efficient exact algorithms (in polynomial complexity ) are not known and may, indeed, even exist

Heurísticas usando construção de vocabuilário aplicadas ao problema da atribuição de localidades a anéis em redes SONET/SDH

The SONET/SDH Ring Assignment Problem (PALAS) treats to group localities in form of some rings, being respected the traffic's limitations of the equipment. Each ring uses a DXC (Digital Cross Connect) to make the communication with the others, being the DXC the equipment most expensive of the net, minimizing the number total of rings, will minimize the total net cost, problem's objective . This topology in rings provides a bigger capacity of regeneration. The PALAS is a problem in Combinatorial Optimization of NP-hard Class. It can be solved through Heuristics and Metaheuristics. In this text, we use Taboo Search while we keep a set of elite solutions to be used in the formation of a part of the collection of vocabulary's parts that in turn will be used in the Vocabulary Building. The Vocabulary Building will be started case Taboo Search does not reach the best solution for the instance. Three approaches had been implemented: one that only uses vocabulary's parts deriving of Taboo Search, one that it only uses vocabulary's parts randomly generated and a last one that it uses half come of the elite and half randomly generated

Gestão do Conhecimento e o uso estratégico de Tecnologias de Informação: um estudo em Redes de Franquias no Estado do Rio Grande do Norte

This study discusses the use of information technologies for knowledge management in networks of franchises in the Rio Grande do Norte/Brazil, whose management and operation are complex activities, characterized by the geographic spread of their network unities, creating barriers to communication and information sharing between franchisors, franchisees and final customers. In view of this, the following hypotheses were formulated: the knowledge management can be a positive alternative for improving communication between units; and information technology can eliminate many problems related mainly to capture and share knowledge. In general, it aims to investigate, in qualitative and quantitative aspects, how information technology can support knowledge management in networks of franchises. Specifically purposes to register the existence of managerial practices related to knowledge management in enterprises at the franchising sector; to verify whether they have the technological resources with the potential to facilitate the sharing of information; to identify what are the technologies of information and communication used in the organizational environment; and suggest measures that will facilitate the process of organizational learning, using information technology and communication as tools. It concludes that knowledge management becomes a positive alternative, especially in strengthening of bonds of communication and sharing of knowledge between the franchises. In this regard, information technology must provide all the services of the corporation to facilitate communication between franchisor and franchisee, through a single and integrated system. However, they still show unsuitable for more sophisticated technology platforms