Learning techniques for automatic email message tagging

A organização automática de mensagens de correio electrónico é um desafio actual na área da aprendizagem automática. O número excessivo de mensagens afecta cada vez mais utilizadores, especialmente os que usam o correio electrónico como ferramenta de comunicação e trabalho. Esta tese aborda o problema da organização automática de mensagens de correio electrónico propondo uma solução que tem como objectivo a etiquetagem automática de mensagens. A etiquetagem automática é feita com recurso às pastas de correio electrónico anteriormente criadas pelos utilizadores, tratando-as como etiquetas, e à sugestão de múltiplas etiquetas para cada mensagem (top-N). São estudadas várias técnicas de aprendizagem e os vários campos que compõe uma mensagem de correio electrónico são analisados de forma a determinar a sua adequação como elementos de classificação. O foco deste trabalho recai sobre os campos textuais (o assunto e o corpo das mensagens), estudando-se diferentes formas de representação, selecção de características e algoritmos de classificação. É ainda efectuada a avaliação dos campos de participantes através de algoritmos de classificação que os representam usando o modelo vectorial ou como um grafo. Os vários campos são combinados para classificação utilizando a técnica de combinação de classificadores Votação por Maioria. Os testes são efectuados com um subconjunto de mensagens de correio electrónico da Enron e um conjunto de dados privados disponibilizados pelo Institute for Systems and Technologies of Information, Control and Communication (INSTICC). Estes conjuntos são analisados de forma a perceber as características dos dados. A avaliação do sistema é realizada através da percentagem de acerto dos classificadores. Os resultados obtidos apresentam melhorias significativas em comparação com os trabalhos relacionados.

What drives corporate default risk premia? Evidence from the CDS market

This paper studies the evolution of the default risk premia for European firms during the years surrounding the recent credit crisis. We employ the information embedded in Credit Default Swaps (CDS) and Moody’s KMV EDF default probabilities to analyze the common factors driving this risk premia. The risk premium is characterized in several directions: Firstly, we perform a panel data analysis to capture the relationship between CDS spreads and actual default probabilities. Secondly, we employ the intensity framework of Jarrow et al. (2005) in order to measure the theoretical effect of risk premium on expected bond returns. Thirdly, we carry out a dynamic panel data to identify the macroeconomic sources of risk premium. Finally, a vector autoregressive model analyzes which proportion of the co-movement is attributable to financial or macro variables. Our estimations report coefficients for risk premium substantially higher than previously referred for US firms and a time varying behavior. A dominant factor explains around 60% of the common movements in risk premia. Additionally, empirical evidence suggests a public-to-private risk transfer between the sovereign CDS spreads and corporate risk premia.

Power converter topologies for wind energy conversion systems: Integrated modeling, control strategy and performance simulation

This paper presents new integrated model for variable-speed wind energy conversion systems, considering a more accurate dynamic of the wind turbine, rotor, generator, power converter and filter. Pulse width modulation by space vector modulation associated with sliding mode is used for controlling the power converters. Also, power factor control is introduced at the output of the power converters. Comprehensive performance simulation studies are carried out with matrix, two-level and multilevel power converter topologies in order to adequately assert the system performance. Conclusions are duly drawn.

Chaos and crises in a model for cooperative hunting: A symbolic dynamics approach

In this work we investigate the population dynamics of cooperative hunting extending the McCann and Yodzis model for a three-species food chain system with a predator, a prey, and a resource species. The new model considers that a given fraction sigma of predators cooperates in prey's hunting, while the rest of the population 1-sigma hunts without cooperation. We use the theory of symbolic dynamics to study the topological entropy and the parameter space ordering of the kneading sequences associated with one-dimensional maps that reproduce significant aspects of the dynamics of the species under several degrees of cooperative hunting. Our model also allows us to investigate the so-called deterministic extinction via chaotic crisis and transient chaos in the framework of cooperative hunting. The symbolic sequences allow us to identify a critical boundary in the parameter spaces (K, C-0) and (K, sigma) which separates two scenarios: (i) all-species coexistence and (ii) predator's extinction via chaotic crisis. We show that the crisis value of the carrying capacity K-c decreases at increasing sigma, indicating that predator's populations with high degree of cooperative hunting are more sensitive to the chaotic crises. We also show that the control method of Dhamala and Lai [Phys. Rev. E 59, 1646 (1999)] can sustain the chaotic behavior after the crisis for systems with cooperative hunting. We finally analyze and quantify the inner structure of the target regions obtained with this control method for wider parameter values beyond the crisis, showing a power law dependence of the extinction transients on such critical parameters.

Two-Higgs-doublet model in light of the standard model H -> tau(+)tau(-) search at the LHC

We study the implications of the searches based on H -> tau(+)tau-by the ATLAS and CMS collaborations on the parameter space of the two-Higgs-doublet model (2HDM). In the 2HDM, the scalars can decay into a tau pair with a branching ratio larger than the SM one, leading to constraints on the 2HDM parameter space. We show that in model II, values of tan beta > 1.8 are definitively excluded if the pseudoscalar is in the mass range 110 GeV < m(A) < 145 GeV. We have also discussed the implications for the 2HDM of the recent dimuon search by the ATLAS collaboration for a CP-odd scalar in the mass range 4-12 GeV.

IOPT Petri Net State Space Generation Algorithm with Maximal-Step Execution Semantics

This paper presents an algorithm to efficiently generate the state-space of systems specified using the IOPT Petri-net modeling formalism. IOPT nets are a non-autonomous Petri-net class, based on Place-Transition nets with an extended set of features designed to allow the rapid prototyping and synthesis of system controllers through an existing hardware-software co-design framework. To obtain coherent and deterministic operation, IOPT nets use a maximal-step execution semantics where, in a single execution step, all enabled transitions will fire simultaneously. This fact increases the resulting state-space complexity and can cause an arc "explosion" effect. Real-world applications, with several million states, will reach a higher order of magnitude number of arcs, leading to the need for high performance state-space generator algorithms. The proposed algorithm applies a compilation approach to read a PNML file containing one IOPT model and automatically generate an optimized C program to calculate the corresponding state-space.

Short-term Feature Space and Music Genre Classification

In music genre classification, most approaches rely on statistical characteristics of low-level features computed on short audio frames. In these methods, it is implicitly considered that frames carry equally relevant information loads and that either individual frames, or distributions thereof, somehow capture the specificities of each genre. In this paper we study the representation space defined by short-term audio features with respect to class boundaries, and compare different processing techniques to partition this space. These partitions are evaluated in terms of accuracy on two genre classification tasks, with several types of classifiers. Experiments show that a randomized and unsupervised partition of the space, used in conjunction with a Markov Model classifier lead to accuracies comparable to the state of the art. We also show that unsupervised partitions of the space tend to create less hubs.

Geometric picture of generalized-CP and Higgs-family transformations in the two-Higgs-doublet model

In the two-Higgs-doublet model (THDM), generalized-CP transformations (phi(i) -> X-ij phi(*)(j) where X is unitary) and unitary Higgs-family transformations (phi(i) -> U-ij phi(j)) have recently been examined in a series of papers. In terms of gauge-invariant bilinear functions of the Higgs fields phi(i), the Higgs-family transformations and the generalized-CP transformations possess a simple geometric description. Namely, these transformations correspond in the space of scalar-field bilinears to proper and improper rotations, respectively. In this formalism, recent results relating generalized CP transformations with Higgs-family transformations have a clear geometric interpretation. We will review what is known regarding THDM symmetries, as well as derive new results concerning those symmetries, namely how they can be interpreted geometrically as applications of several CP transformations.

Topological Complexity and Predictability in the Dynamics of a Tumor Growth Model with Shilnikov's Chaos

Dynamical systems modeling tumor growth have been investigated to determine the dynamics between tumor and healthy cells. Recent theoretical investigations indicate that these interactions may lead to different dynamical outcomes, in particular to homoclinic chaos. In the present study, we analyze both topological and dynamical properties of a recently characterized chaotic attractor governing the dynamics of tumor cells interacting with healthy tissue cells and effector cells of the immune system. By using the theory of symbolic dynamics, we first characterize the topological entropy and the parameter space ordering of kneading sequences from one-dimensional iterated maps identified in the dynamics, focusing on the effects of inactivation interactions between both effector and tumor cells. The previous analyses are complemented with the computation of the spectrum of Lyapunov exponents, the fractal dimension and the predictability of the chaotic attractors. Our results show that the inactivation rate of effector cells by the tumor cells has an important effect on the dynamics of the system. The increase of effector cells inactivation involves an inverse Feigenbaum (i.e. period-halving bifurcation) scenario, which results in the stabilization of the dynamics and in an increase of dynamics predictability. Our analyses also reveal that, at low inactivation rates of effector cells, tumor cells undergo strong, chaotic fluctuations, with the dynamics being highly unpredictable. Our findings are discussed in the context of tumor cells potential viability.

Portuguese newspapers’ front-page and the coverage of European Cultural Capital: Porto 2001 and Guimarães 2012

Portugal hosted in the last thirteen years, two editions of the event European Cultural Capital; this paper intends to illustrate the coverage that Portuguese newspapers (daily newspapers Público, Diário de Notícias, Correio da Manhã and Jornal de Notícias, a weekly newsmagazine Visão and a weekly newspaper Expresso) made, through referrals in front-page and respective developments within the editions, to each of the events and that allows us to define the main moments that marked each of them, patterns of action, the major players, planning and programming types. The European Cultural Capital project elects, from year to year, cities of different EU member states with the main goal of “contributing to bring together the Europe´s people" (words of Mélina Mercouri, Greek Minister of Culture who, in 1985, proposed the launch of this initiative) and encouraging the elected urban space to present new cultural paradigms. In the genesis of this model is the cultural decentralization’s vector, a possibility to medium-sized cities of funding public works, restoring heritage and promoting themselves in touristic terms, of giving visibility to cities away from cultural and creative industries’ major distribution centers. A crucial factor to achieve this goal is media coverage. This paper outline the information that the Portuguese press ran over the two years that elapsed the latest editions of the European Cultural Capital in Portugal, namely that media coverage have deviated from the disclosure of the events’ schedule to suggest itineraries of visit and little or not even question the role that cities, promoting such initiatives, have as places of innovation in terms of cultural policies, artistic production and innovation, in urban and environmental regeneration, in economic revitalization, in training and creating new artists and new audiences and in boosting the confidence of local communities. The content analysis performed to articles shows how press is essential to the promotion of cities as cultural/touristic destinations as it stimulates consumption among residents and attracts visitors, with the possible dire consequence of turning the cultural journalist into an agent of touristic instead of cultural promotion.

Newspaper coverage impact on cultural events: Porto 2001 and Guimarães 2012: European Cultural Capitals

Artigo baseado na comunicação proferida no 1st International Symposium on Media Studies, realizado na Akdeniz Universitesi Yayınları, Antalya, Turquia, 21-23 de novembro de 2013

Simulation of offshore wind system with two-level converters: HVDC power transmission

A new integrated mathematical model for the simulation of offshore wind energy conversion system performance is presented in this paper. The mathematical model considers an offshore variable-speed turbine in deep water equipped with a permanent magnet synchronous generator using full-power two-level converter, converting the energy of a variable frequency source in injected energy into the electric network with constant frequency, through a high voltage DC transmission submarine cable. The mathematical model for the drive train is a concentrate two mass model which incorporates the dynamic for the structure and tower due to the need to emulate the effects of the moving surface. Controller strategy considered is a proportional integral one. Also, pulse width modulation using space vector modulation supplemented with sliding mode is used for trigger the transistor of the converter. Finally, a case study is presented to access the system performance. © 2014 IEEE.

Simulation of offshore wind system with two levels converters: HVDC power transmission

A new integrated mathematical model for the simulation of offshore wind energy conversion system performance is presented in this paper. The mathematical model considers an offshore variable-speed turbine in deep water equipped with a permanent magnet synchronous generator using full-power two-level converter, converting the energy of a variable frequency source in injected energy into the electric network with constant frequency, through a high voltage DC transmission submarine cable. The mathematical model for the drive train is a concentrate two mass model which incorporates the dynamic for the structure and tower due to the need to emulate the effects of the moving surface. Controller strategy considered is a proportional integral one. Also, pulse width modulation using space vector modulation supplemented with sliding mode is used for trigger the transistor of the converter. Finally, a case study is presented to access the system performance. © 2014 IEEE.

Preserving the validity of the two-Higgs-doublet model up to the Planck scale

We examine the constraints on the two Higgs doublet model (2HDM) due to the stability of the scalar potential and absence of Landau poles at energy scales below the Planck scale. We employ the most general 2HDM that incorporates an approximately Standard Model (SM) Higgs boson with a flavor aligned Yukawa sector to eliminate potential tree-level Higgs-mediated flavor changing neutral currents. Using basis independent techniques, we exhibit robust regimes of the 2HDM parameter space with a 125 GeV SM-like Higgs boson that is stable and perturbative up to the Planck scale. Implications for the heavy scalar spectrum are exhibited.

Vertex component analysis: a geometric-based approach to unmix hyperspectral data

Hyperspectral remote sensing exploits the electromagnetic scattering patterns of the different materials at specific wavelengths [2, 3]. Hyperspectral sensors have been developed to sample the scattered portion of the electromagnetic spectrum extending from the visible region through the near-infrared and mid-infrared, in hundreds of narrow contiguous bands [4, 5]. The number and variety of potential civilian and military applications of hyperspectral remote sensing is enormous [6, 7]. Very often, the resolution cell corresponding to a single pixel in an image contains several substances (endmembers) [4]. In this situation, the scattered energy is a mixing of the endmember spectra. A challenging task underlying many hyperspectral imagery applications is then decomposing a mixed pixel into a collection of reflectance spectra, called endmember signatures, and the corresponding abundance fractions [8–10]. Depending on the mixing scales at each pixel, the observed mixture is either linear or nonlinear [11, 12]. Linear mixing model holds approximately when the mixing scale is macroscopic [13] and there is negligible interaction among distinct endmembers [3, 14]. If, however, the mixing scale is microscopic (or intimate mixtures) [15, 16] and the incident solar radiation is scattered by the scene through multiple bounces involving several endmembers [17], the linear model is no longer accurate. Linear spectral unmixing has been intensively researched in the last years [9, 10, 12, 18–21]. It considers that a mixed pixel is a linear combination of endmember signatures weighted by the correspondent abundance fractions. Under this model, and assuming that the number of substances and their reflectance spectra are known, hyperspectral unmixing is a linear problem for which many solutions have been proposed (e.g., maximum likelihood estimation [8], spectral signature matching [22], spectral angle mapper [23], subspace projection methods [24,25], and constrained least squares [26]). In most cases, the number of substances and their reflectances are not known and, then, hyperspectral unmixing falls into the class of blind source separation problems [27]. Independent component analysis (ICA) has recently been proposed as a tool to blindly unmix hyperspectral data [28–31]. ICA is based on the assumption of mutually independent sources (abundance fractions), which is not the case of hyperspectral data, since the sum of abundance fractions is constant, implying statistical dependence among them. This dependence compromises ICA applicability to hyperspectral images as shown in Refs. [21, 32]. In fact, ICA finds the endmember signatures by multiplying the spectral vectors with an unmixing matrix, which minimizes the mutual information among sources. If sources are independent, ICA provides the correct unmixing, since the minimum of the mutual information is obtained only when sources are independent. This is no longer true for dependent abundance fractions. Nevertheless, some endmembers may be approximately unmixed. These aspects are addressed in Ref. [33]. Under the linear mixing model, the observations from a scene are in a simplex whose vertices correspond to the endmembers. Several approaches [34–36] have exploited this geometric feature of hyperspectral mixtures [35]. Minimum volume transform (MVT) algorithm [36] determines the simplex of minimum volume containing the data. The method presented in Ref. [37] is also of MVT type but, by introducing the notion of bundles, it takes into account the endmember variability usually present in hyperspectral mixtures. The MVT type approaches are complex from the computational point of view. Usually, these algorithms find in the first place the convex hull defined by the observed data and then fit a minimum volume simplex to it. For example, the gift wrapping algorithm [38] computes the convex hull of n data points in a d-dimensional space with a computational complexity of O(nbd=2cþ1), where bxc is the highest integer lower or equal than x and n is the number of samples. The complexity of the method presented in Ref. [37] is even higher, since the temperature of the simulated annealing algorithm used shall follow a log( ) law [39] to assure convergence (in probability) to the desired solution. Aiming at a lower computational complexity, some algorithms such as the pixel purity index (PPI) [35] and the N-FINDR [40] still find the minimum volume simplex containing the data cloud, but they assume the presence of at least one pure pixel of each endmember in the data. This is a strong requisite that may not hold in some data sets. In any case, these algorithms find the set of most pure pixels in the data. PPI algorithm uses the minimum noise fraction (MNF) [41] as a preprocessing step to reduce dimensionality and to improve the signal-to-noise ratio (SNR). The algorithm then projects every spectral vector onto skewers (large number of random vectors) [35, 42,43]. The points corresponding to extremes, for each skewer direction, are stored. A cumulative account records the number of times each pixel (i.e., a given spectral vector) is found to be an extreme. The pixels with the highest scores are the purest ones. N-FINDR algorithm [40] is based on the fact that in p spectral dimensions, the p-volume defined by a simplex formed by the purest pixels is larger than any other volume defined by any other combination of pixels. This algorithm finds the set of pixels defining the largest volume by inflating a simplex inside the data. ORA SIS [44, 45] is a hyperspectral framework developed by the U.S. Naval Research Laboratory consisting of several algorithms organized in six modules: exemplar selector, adaptative learner, demixer, knowledge base or spectral library, and spatial postrocessor. The first step consists in flat-fielding the spectra. Next, the exemplar selection module is used to select spectral vectors that best represent the smaller convex cone containing the data. The other pixels are rejected when the spectral angle distance (SAD) is less than a given thresh old. The procedure finds the basis for a subspace of a lower dimension using a modified Gram–Schmidt orthogonalizati on. The selected vectors are then projected onto this subspace and a simplex is found by an MV T pro cess. ORA SIS is oriented to real-time target detection from uncrewed air vehicles using hyperspectral data [46]. In this chapter we develop a new algorithm to unmix linear mixtures of endmember spectra. First, the algorithm determines the number of endmembers and the signal subspace using a newly developed concept [47, 48]. Second, the algorithm extracts the most pure pixels present in the data. Unlike other methods, this algorithm is completely automatic and unsupervised. To estimate the number of endmembers and the signal subspace in hyperspectral linear mixtures, the proposed scheme begins by estimating sign al and noise correlation matrices. The latter is based on multiple regression theory. The signal subspace is then identified by selectin g the set of signal eigenvalue s that best represents the data, in the least-square sense [48,49 ], we note, however, that VCA works with projected and with unprojected data. The extraction of the end members exploits two facts: (1) the endmembers are the vertices of a simplex and (2) the affine transformation of a simplex is also a simplex. As PPI and N-FIND R algorithms, VCA also assumes the presence of pure pixels in the data. The algorithm iteratively projects data on to a direction orthogonal to the subspace spanned by the endmembers already determined. The new end member signature corresponds to the extreme of the projection. The algorithm iterates until all end members are exhausted. VCA performs much better than PPI and better than or comparable to N-FI NDR; yet it has a computational complexity between on e and two orders of magnitude lower than N-FINDR. The chapter is structure d as follows. Section 19.2 describes the fundamentals of the proposed method. Section 19.3 and Section 19.4 evaluate the proposed algorithm using simulated and real data, respectively. Section 19.5 presents some concluding remarks.