Statistical Tests for Joint Analysis of Performance Measures

Recently there has been an increasing interest in the development of new methods using Pareto optimality to deal with multi-objective criteria (for example, accuracy and architectural complexity). Once one has learned a model based on their devised method, the problem is then how to compare it with the state of art. In machine learning, algorithms are typically evaluated by comparing their performance on different data sets by means of statistical tests. Unfortunately, the standard tests used for this purpose are not able to jointly consider performance measures. The aim of this paper is to resolve this issue by developing statistical procedures that are able to account for multiple competing measures at the same time. In particular, we develop two tests: a frequentist procedure based on the generalized likelihood-ratio test and a Bayesian procedure based on a multinomial-Dirichlet conjugate model. We further extend them by discovering conditional independences among measures to reduce the number of parameter of such models, as usually the number of studied cases is very reduced in such comparisons. Real data from a comparison among general purpose classifiers is used to show a practical application of our tests.

Dimensioning of multilayer optical networks

Este trabalho apresenta um estudo sobre o dimensionamento de redes ópticas, com vistas a obter um modelo de dimensionamento para redes de transporte sobreviventes. No estudo utilizou-se uma abordagem estatística em detrimento à determinística. Inicialmente, apresentam-se as principais tecnologias e diferentes arquitecturas utilizadas nas redes ópticas de transporte. Bem como os principais esquemas de sobrevivência e modos de transporte. São identificadas variáveis necessárias e apresenta-se um modelo dimensionamento para redes de transporte, tendo-se dado ênfase às redes com topologia em malha e considerando os modos de transporte opaco, transparente e translúcido. É feita uma análise rigorosa das características das topologias de redes de transporte reais, e desenvolve-se um gerador de topologias de redes de transporte, para testar a validade dos modelos desenvolvidos. Também é implementado um algoritmo genético para a obtenção de uma topologia optimizada para um dado tráfego. São propostas expressões para o cálculo de variáveis não determinísticas, nomeadamente, para o número médio de saltos de um pedido, coeficiente de protecção e coeficiente de restauro. Para as duas últimas, também é analisado o impacto do modelo de tráfego. Verifica-se que os resultados obtidos pelas expressões propostas são similares às obtidas por cálculo numérico, e que o modelo de tráfego não influencia significativamente os valores obtidos para os coeficientes. Finalmente, é demonstrado que o modelo proposto é útil para o dimensionamento e cálculo dos custos de capital de redes com informação incompleta.

Methodologies for traffic profiling in communication networks

Internet Tra c, Internet Applications, Internet Attacks, Tra c Pro ling, Multi-Scale Analysis abstract Nowadays, the Internet can be seen as an ever-changing platform where new and di erent types of services and applications are constantly emerging. In fact, many of the existing dominant applications, such as social networks, have appeared recently, being rapidly adopted by the user community. All these new applications required the implementation of novel communication protocols that present di erent network requirements, according to the service they deploy. All this diversity and novelty has lead to an increasing need of accurately pro ling Internet users, by mapping their tra c to the originating application, in order to improve many network management tasks such as resources optimization, network performance, service personalization and security. However, accurately mapping tra c to its originating application is a di cult task due to the inherent complexity of existing network protocols and to several restrictions that prevent the analysis of the contents of the generated tra c. In fact, many technologies, such as tra c encryption, are widely deployed to assure and protect the con dentiality and integrity of communications over the Internet. On the other hand, many legal constraints also forbid the analysis of the clients' tra c in order to protect their con dentiality and privacy. Consequently, novel tra c discrimination methodologies are necessary for an accurate tra c classi cation and user pro ling. This thesis proposes several identi cation methodologies for an accurate Internet tra c pro ling while coping with the di erent mentioned restrictions and with the existing encryption techniques. By analyzing the several frequency components present in the captured tra c and inferring the presence of the di erent network and user related events, the proposed approaches are able to create a pro le for each one of the analyzed Internet applications. The use of several probabilistic models will allow the accurate association of the analyzed tra c to the corresponding application. Several enhancements will also be proposed in order to allow the identi cation of hidden illicit patterns and the real-time classi cation of captured tra c. In addition, a new network management paradigm for wired and wireless networks will be proposed. The analysis of the layer 2 tra c metrics and the di erent frequency components that are present in the captured tra c allows an e cient user pro ling in terms of the used web-application. Finally, some usage scenarios for these methodologies will be presented and discussed.

Design and optimization next generation passive all-optical networks

Este trabalho investiga novas metodologias para as redes óticas de acesso de próxima geração (NG-OAN). O trabalho está dividido em quatro tópicos de investigação: projeto da rede, modelos numéricos para efeitos não lineares da fibra ótica, impacto dos efeitos não lineares da fibra ótica e otimização da rede. A rede ótica de acesso investigada nesse trabalho está projetado para suprir os requisitos de densidade de utilizadores e cobertura, isto é, suportar muitos utilizadores ( 1000) com altas velocidades de conexão dedicada ( 1 Gb/s) ocupando uma faixa estreita do espectro ( 25 nm) e comprimentos de fibra ótica até 100 km. Os cenários são baseados em redes óticas passivas com multiplexagem por divisão no comprimento de onda de alta densidade (UDWDM-PON) utilizando transmissores/receptores coerentes nos terminais da rede. A rede é avaliada para vários ritmos de transmissão usando formatos de modulação avançados, requisitos de largura de banda por utilizador e partilha de banda com tecnologias tradicionais de redes óticas passivas (PON). Modelos numéricos baseados em funções de transferência das séries de Volterra (VSTF) são demonstrados tanto para a análise dos efeitos não lineares da fibra ótica quanto para avaliação do desempenho total da rede. São apresentadas as faixas de potência e distância de transmissão nas quais as séries de Volterra apresentam resultados semelhantes ao modelo referência Split-Step Fourier (SSF) (validado experimentalmente) para o desempenho total da rede. Além disso, um algoritmo, que evita componentes espectrais com intensidade nulo, é proposto para realizar cálculos rápidos das séries. O modelo VSTF é estendido para identificar unicamente os efeitos não lineares da fibra ótica mais relevantes no cenário investigado: Self-Phase Modulation (SPM), Cross-Phase Modulation (XPM) e Four-Wave Mixing (FWM). Simulações numéricas são apresentadas para identificar o impacto isolado de cada efeito não linear da fibra ótica, SPM, XPM e FWM, no desempenho da rede com detecção coerente UDWDM-PON, transportando canais com modulação digital em fase (M-ária PSK) ou modulação digital em amplitude (M-ária QAM). A análise numérica é estendida para diferentes comprimentos de fibra ótica mono modo (SSMF), potência por canal e ritmo de transmissão por canal. Por conseguinte, expressões analíticas são extrapoladas para determinar a evolução do SPM, XPM e FWM em função da potência e distância de transmissão em cenários NG-OAN. O desempenho da rede é otimizada através da minimização parcial da interferência FWM (via espaçamento desigual dos canais), que nesse caso, é o efeito não linear da fibra ótica mais relevante. Direções para melhorias adicionas no desempenho da rede são apresentados para cenários em que o XPM é relevante, isto é, redes transportando formatos de modulação QAM. A solução, nesse caso, é baseada na utilização de técnicas de processamento digital do sinal.

Analysis, modeling, design and optimization of future communications systems: from theory to practice

Esta tese descreve uma framework de trabalho assente no paradigma multi-camada para analisar, modelar, projectar e optimizar sistemas de comunicação. Nela se explora uma nova perspectiva acerca da camada física que nasce das relações entre a teoria de informação, estimação, métodos probabilísticos, teoria da comunicação e codificação. Esta framework conduz a métodos de projecto para a próxima geração de sistemas de comunicação de alto débito. Além disso, a tese explora várias técnicas de camada de acesso com base na relação entre atraso e débito para o projeto de redes sem fio tolerantes a atrasos. Alguns resultados fundamentais sobre a interação entre a teoria da informação e teoria da estimação conduzem a propostas de um paradigma alternativo para a análise, projecto e optimização de sistemas de comunicação. Com base em estudos sobre a relação entre a informação recíproca e MMSE, a abordagem descrita na tese permite ultrapassar, de forma inovadora, as dificuldades inerentes à optimização das taxas de transmissão de informação confiáveis em sistemas de comunicação, e permite a exploração da atribuição óptima de potência e estruturas óptimas de pre-codificação para diferentes modelos de canal: com fios, sem fios e ópticos. A tese aborda também o problema do atraso, numa tentativa de responder a questões levantadas pela enorme procura de débitos elevados em sistemas de comunicação. Isso é feito através da proposta de novos modelos para sistemas com codificação de rede (network coding) em camadas acima da sua camada física. Em particular, aborda-se a utilização de sistemas de codificação em rede para canais que variam no tempo e são sensíveis a atrasos. Isso foi demonstrado através da proposta de um novo modelo e esquema adaptativo, cujos algoritmos foram aplicados a sistemas sem fios com desvanecimento (fading) complexo, de que são exemplos os sistemas de comunicação via satélite. A tese aborda ainda o uso de sistemas de codificação de rede em cenários de transferência (handover) exigentes. Isso é feito através da proposta de novos modelos de transmissão WiFi IEEE 801.11 MAC, que são comparados com codificação de rede, e que se demonstram possibilitar transferência sem descontinuidades. Pode assim dizer-se que esta tese, através de trabalho de análise e de propostas suportadas por simulações, defende que na concepção de sistemas de comunicação se devem considerar estratégias de transmissão e codificação que sejam não só próximas da capacidade dos canais, mas também tolerantes a atrasos, e que tais estratégias têm de ser concebidas tendo em vista características do canal e a camada física.

Monitoring and all-optical signal processing techniques for transparent networks

Optical networks are under constant evolution. The growing demand for dynamism require devices that can accommodate different types of traffic. Thus the study of transparent optical networks arises. This approach makes optical networks more "elegant" , due to a more efficient use of network resources. In this thesis, the author proposes devices that intend to form alternative approaches both in the state of art of these same technologies both in the fitting of this technologies in transparent optical networks. Given that full transparency is difficult to achieve with current technology (perhaps with more developed optical computing this is possible), the author proposes techniques with different levels of transparency. On the topic of performance of optical networks, the author proposes two techniques for monitoring chromatic dispersion with different levels of transparency. In Chapter 3 the proposed technique seems to make more sense for long-haul optical transmission links and high transmission rates, not only due to its moderate complexity but also to its potential moderate/high cost. However it is proposed to several modulation formats, particularly those that have a protruding clock component. In Chapter 4 the transparency level was not tested for various modulation formats, however some transparency is achieved by not adding any electrical device after the receiver (other than an analog-digital converter). This allows that this technique can operate at high transmission rates in excess of 100 Gbit / s, if electro-optical asynchronous sampling is used before the optical receiver. Thus a low cost and low bandwidth photo-detector can be used. In chapter 5 is demonstrated a technique for simultaneously monitoring multiple impairments of the optical network by generating novel performance analysis diagrams and by use of artificial neural networks. In chapter 6 the author demonstrates an all-optical technique for controlling the optical state of polarization and an example of how all-optical signal processing can fully cooperate with optical performance monitoring.

Genetic programming and bacterial algorithm for neural networks and fuzzy systems design

In the field of control systems it is common to use techniques based on model adaptation to carry out control for plants for which mathematical analysis may be intricate. Increasing interest in biologically inspired learning algorithms for control techniques such as Artificial Neural Networks and Fuzzy Systems is in progress. In this line, this paper gives a perspective on the quality of results given by two different biologically connected learning algorithms for the design of B-spline neural networks (BNN) and fuzzy systems (FS). One approach used is the Genetic Programming (GP) for BNN design and the other is the Bacterial Evolutionary Algorithm (BEA) applied for fuzzy rule extraction. Also, the facility to incorporate a multi-objective approach to the GP algorithm is outlined, enabling the designer to obtain models more adequate for their intended use.

Evaluation of the H.264 scalable video coding in error prone IP networks

The Joint Video Team, composed by the ISO/IEC Moving Picture Experts Group (MPEG) and the ITU-T Video Coding Experts Group (VCEG), has standardized a scalable extension of the H.264/AVC video coding standard called Scalable Video Coding (SVC). H.264/SVC provides scalable video streams which are composed by a base layer and one or more enhancement layers. Enhancement layers may improve the temporal, the spatial or the signal-to-noise ratio resolutions of the content represented by the lower layers. One of the applications, of this standard is related to video transmission in both wired and wireless communication systems, and it is therefore important to analyze in which way packet losses contribute to the degradation of quality, and which mechanisms could be used to improve that quality. This paper provides an analysis and evaluation of H.264/SVC in error prone environments, quantifying the degradation caused by packet losses in the decoded video. It also proposes and analyzes the consequences of QoS-based discarding of packets through different marking solutions.

Monitoring of urban transportation networks using wireless sensor networks

Dissertação de mestrado, Engenharia Informática, Faculdade de Ciências e Tecnologia, Universidade do Algarve, 2015

Adaptive Probabilistic Topic Models for Social Networks

Thesis (Master's)--University of Washington, 2012

Design and Qualitative/Quantitative Analysis of Multi-Agent Spatial Simulation Library

Thesis (Master's)--University of Washington, 2012

PREGNANCY-RELATED CHURN AND ACCESS TO CARE IN THE CONTEXT OF NARROW NETWORKS IN WASHINGTON STATE

Thesis (Master's)--University of Washington, 2016-03

Specifying and analysing networks of processes in CSPt (or in search of associativity)

In proposing theories of how we should design and specify networks of processes it is necessary to show that the semantics of any language we use to write down the intended behaviours of a system has several qualities. First in that the meaning of what is written on the page reflects the intention of the designer; second that there are no unexpected behaviours that might arise in a specified system that are hidden from the unsuspecting specifier; and third that the intention for the design of the behaviour of a network of processes can be communicated clearly and intuitively to others. In order to achieve this we have developed a variant of CSP, called CSPt, designed to solve the problems of termination of parallel processes present in the original formulation of CSP. In CSPt we introduced three parallel operators, each with a different kind of termination semantics, which we call synchronous, asynchronous and race. These operators provide specifiers with an expressive and flexible tool kit to define the intended behaviour of a system in such a way that unexpected or unwanted behaviours are guaranteed not to take place. In this paper we extend out analysis of CSPt and introduce the notion of an alphabet diagram that illustrates the different categories of events that can arise in the parallel composition of processes. These alphabet diagrams are then used to analyse networks of three processes in parallel with the aim of identifying sufficient constraints to ensure associativity of their parallel composition. Having achieved this we then proceed to prove associativity laws for the three parallel operators of CSPt. Next, we illustrate how to design and construct a network of three processes that satisfy the associativity law, using the associativity theorem and alphabet diagrams. Finally, we outline how this could be achieved for more general networks of processes.

Identification of meat spoilage by FTIR spectroscopy and neural networks

Freshness and safety of muscle foods are generally considered as the most important parameters for the food industry. To address the rapid determination of meat spoilage, Fourier transform infrared (FTIR) spectroscopy technique, with the help of advanced learning-based methods, was attempted in this work. FTIR spectra were obtained from the surface of beef samples during aerobic storage at various temperatures, while a microbiological analysis had identified the population of Total viable counts. A fuzzy principal component algorithm has been also developed to reduce the dimensionality of the spectral data. The results confirmed the superiority of the adopted scheme compared to the partial least squares technique, currently used in food microbiology.

Zonal prices analysis supported by a data mining based methodology

A methodology based on data mining techniques to support the analysis of zonal prices in real transmission networks is proposed in this paper. The mentioned methodology uses clustering algorithms to group the buses in typical classes that include a set of buses with similar LMP values. Two different clustering algorithms have been used to determine the LMP clusters: the two-step and K-means algorithms. In order to evaluate the quality of the partition as well as the best performance algorithm adequacy measurements indices are used. The paper includes a case study using a Locational Marginal Prices (LMP) data base from the California ISO (CAISO) in order to identify zonal prices.