Evolutionary tuning of SVM parameter values in multiclass problems

Support vector machines (SVMs) were originally formulated for the solution of binary classification problems. In multiclass problems, a decomposition approach is often employed, in which the multiclass problem is divided into multiple binary subproblems, whose results are combined. Generally, the performance of SVM classifiers is affected by the selection of values for their parameters. This paper investigates the use of genetic algorithms (GAs) to tune the parameters of the binary SVMs in common multiclass decompositions. The developed GA may search for a set of parameter values common to all binary classifiers or for differentiated values for each binary classifier. (C) 2008 Elsevier B.V. All rights reserved.

A hybrid case adaptation approach for case-based reasoning

Case-Based Reasoning is a methodology for problem solving based on past experiences. This methodology tries to solve a new problem by retrieving and adapting previously known solutions of similar problems. However, retrieved solutions, in general, require adaptations in order to be applied to new contexts. One of the major challenges in Case-Based Reasoning is the development of an efficient methodology for case adaptation. The most widely used form of adaptation employs hand coded adaptation rules, which demands a significant knowledge acquisition and engineering effort. An alternative to overcome the difficulties associated with the acquisition of knowledge for case adaptation has been the use of hybrid approaches and automatic learning algorithms for the acquisition of the knowledge used for the adaptation. We investigate the use of hybrid approaches for case adaptation employing Machine Learning algorithms. The approaches investigated how to automatically learn adaptation knowledge from a case base and apply it to adapt retrieved solutions. In order to verify the potential of the proposed approaches, they are experimentally compared with individual Machine Learning techniques. The results obtained indicate the potential of these approaches as an efficient approach for acquiring case adaptation knowledge. They show that the combination of Instance-Based Learning and Inductive Learning paradigms and the use of a data set of adaptation patterns yield adaptations of the retrieved solutions with high predictive accuracy.

HIERARCHICAL DECOMPOSITION OF MULTICLASS PROBLEMS

Several popular Machine Learning techniques are originally designed for the solution of two-class problems. However, several classification problems have more than two classes. One approach to deal with multiclass problems using binary classifiers is to decompose the multiclass problem into multiple binary sub-problems disposed in a binary tree. This approach requires a binary partition of the classes for each node of the tree, which defines the tree structure. This paper presents two algorithms to determine the tree structure taking into account information collected from the used dataset. This approach allows the tree structure to be determined automatically for any multiclass dataset.

Comparing machine learning classifiers in potential distribution modelling

Species` potential distribution modelling consists of building a representation of the fundamental ecological requirements of a species from biotic and abiotic conditions where the species is known to occur. Such models can be valuable tools to understand the biogeography of species and to support the prediction of its presence/absence considering a particular environment scenario. This paper investigates the use of different supervised machine learning techniques to model the potential distribution of 35 plant species from Latin America. Each technique was able to extract a different representation of the relations between the environmental conditions and the distribution profile of the species. The experimental results highlight the good performance of random trees classifiers, indicating this particular technique as a promising candidate for modelling species` potential distribution. (C) 2010 Elsevier Ltd. All rights reserved.

A hybrid approach to learn with imbalanced classes using evolutionary algorithms

There is an increasing interest in the application of Evolutionary Algorithms (EAs) to induce classification rules. This hybrid approach can benefit areas where classical methods for rule induction have not been very successful. One example is the induction of classification rules in imbalanced domains. Imbalanced data occur when one or more classes heavily outnumber other classes. Frequently, classical machine learning (ML) classifiers are not able to learn in the presence of imbalanced data sets, inducing classification models that always predict the most numerous classes. In this work, we propose a novel hybrid approach to deal with this problem. We create several balanced data sets with all minority class cases and a random sample of majority class cases. These balanced data sets are fed to classical ML systems that produce rule sets. The rule sets are combined creating a pool of rules and an EA is used to build a classifier from this pool of rules. This hybrid approach has some advantages over undersampling, since it reduces the amount of discarded information, and some advantages over oversampling, since it avoids overfitting. The proposed approach was experimentally analysed and the experimental results show an improvement in the classification performance measured as the area under the receiver operating characteristics (ROC) curve.

Positive solutions for a fourth order equation with nonlinear boundary conditions

Existence of positive solutions for a fourth order equation with nonlinear boundary conditions, which models deformations of beams on elastic supports, is considered using fixed points theorems in cones of ordered Banach spaces. Iterative and numerical solutions are also considered. (C) 2010 IMACS. Published by Elsevier B.V. All rights reserved.

A collocation method for solving singular integro-differential equations

This paper describes a collocation method for numerically solving Cauchy-type linear singular integro-differential equations. The numerical method is based on the transformation of the integro-differential equation into an integral equation, and then applying a collocation method to solve the latter. The collocation points are chosen as the Chebyshev nodes. Uniform convergence of the resulting method is then discussed. Numerical examples are presented and solved by the numerical techniques.

Influence of quantum-dots density on average in-plane strain of optoelectronic devices investigated by high-resolution X-ray diffraction

High-resolution X-ray diffractometry is used to probe the nature of a diffraction-peak broadening previously noticed in quantum dots (QDs) systems with freestanding InAs islands on top of GaAs (001) substrates [Freitas et al., Phys. Status Solidi (A) 204, 2548 (2007)]. The procedure is hence extended to further investigate the capping process of InAs/GaAs QDs. A direct correlation is established between QDs growth rates and misorientation of lattice-planes at the samples surfaces. This effect provides an alternative too] for studying average strain fields on QDs systems in standard triple axis diffractometers running on X-ray tube sources, which are much more common than synchrotron facilities. (C) 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Exciton behavior in GaAs/AlGaAs coupled double quantum wells with interface disorder

In this work, we present a detailed study on the optical properties of two GaAs/Al(0.35)Ga(0.65)As coupled double quantum wells (CDQWs) with inter-well barriers of different thicknesses, by using photoluminescence (PL) spectroscopy. The two CDQWs were grown in a single sample, assuring very similar experimental conditions for measurements of both. The PL spectrum of each CDQW exhibits two recombination channels which can be accurately identified as the excitonic e(1)-hh(1) transitions originated from CDQWs of different effective dimensions. The PL spectra characteristics and the behavior of the emissions as a function of temperature and excitation power are interpreted in the scenario of the bimodal interface roughness model, taking into account the exciton migration between the two regions considered in this model and the difference in the potential fluctuation levels between those two regions. The details of the PL spectra behavior as a function of excitation power are explained in terms of the competition between the band gap renormalization (BGR) and the potential fluctuation effects. The results obtained for the two CDQWs, which have different degrees of potential fluctuation, are also compared and discussed. (C) 2009 Elsevier B.V. All rights reserved.

Opinion dynamics of learning agents: does seeking consensus lead to disagreement?

We study opinion dynamics in a population of interacting adaptive agents voting on a set of issues represented by vectors. We consider agents who can classify issues into one of two categories and can arrive at their opinions using an adaptive algorithm. Adaptation comes from learning and the information for the learning process comes from interacting with other neighboring agents and trying to change the internal state in order to concur with their opinions. The change in the internal state is driven by the information contained in the issue and in the opinion of the other agent. We present results in a simple yet rich context where each agent uses a Boolean perceptron to state their opinion. If the update occurs with information asynchronously exchanged among pairs of agents, then the typical case, if the number of issues is kept small, is the evolution into a society torn by the emergence of factions with extreme opposite beliefs. This occurs even when seeking consensus with agents with opposite opinions. If the number of issues is large, the dynamics becomes trapped, the society does not evolve into factions and a distribution of moderate opinions is observed. The synchronous case is technically simpler and is studied by formulating the problem in terms of differential equations that describe the evolution of order parameters that measure the consensus between pairs of agents. We show that for a large number of issues and unidirectional information flow, global consensus is a fixed point; however, the approach to this consensus is glassy for large societies.

The effect of confinement on the temperature dependence of the excitonic transition energy in GaAs/Al(x)Ga(1-x)As quantum wells

We determined by means of photoluminescence measurements the dependence on temperature of the transition energy of excitons in GaAs/Al(x)Ga(1-x)As quantum wells with different alloy concentrations (with different barrier heights). Using a fitting procedure, we determined the parameters which describe the behavior of the excitonic transition energy as a function of temperature according to three different theoretical models. We verified that the temperature dependence of the excitonic transition energy does not only depend on the GaAs material but also depends on the barrier material, i.e. on the alloy composition. The effect of confinement on the temperature dependence of the excitonic transition is discussed.

Growth of EuTe islands on SnTe by molecular beam epitaxy

Semiconductor magnetic quantum dots are very promising structures, with novel properties that find multiple applications in spintronic devices. EuTe is a wide gap semiconductor with NaCl structure, and strong magnetic moments S=7/2 at the half filled 4f(7) electronic levels. On the other hand, SnTe is a narrow gap semiconductor with the same crystal structure and 4% lattice mismatch with EuTe. In this work, we investigate the molecular beam epitaxial growth of EuTe on SnTe after the critical thickness for island formation is surpassed, as a previous step to the growth of organized magnetic quantum dots. The topology and strain state of EuTe islands were studied as a function of growth temperature and EuTe nominal layer thickness. Reflection high energy electron diffraction (RHEED) was used in-situ to monitor surface morphology and strain state. RHEED results were complemented and enriched with atomic force microscopy and grazing incidence X-ray diffraction measurements made at the XRD2 beamline of the Brazilian Synchrotron. EuTe islands of increasing height and diameter are obtained when the EuTe nominal thickness increases, with higher aspect ratio for the islands grown at lower temperatures. As the islands grow, a relaxation toward the EuTe bulk lattice parameter was observed. The relaxation process was partially reverted by the growth of the SnTe cap layer, vital to protect the EuTe islands from oxidation. A simple model is outlined to describe the distortions caused by the EuTe islands on the SnTe buffer and cap layers. The SnTe cap layers formed interesting plateau structures with easily controlled wall height, that could find applications as a template for future nanostructures growth. (C) 2010 Elsevier B.V. All rights reserved.

MAGNETO-OPTICAL ABSORPTION AND PHOTOMAGNETISM IN EUROPIUM CHALCOGENIDES

The absorption threshold in EuTe and EuSe was investigated as a function of applied magnetic field in the Faraday geometry. A well-resolved doublet of sharp dichroic lines was observed when the magnetic field induced ferromagnetic alignment of the spins in the crystal lattice. In contrast, at zero magnetic field only a broad and featureless absorption onset is seen. These results are fully explained in terms of a model of electronic transitions between localized states at the Eu lattice site and a tight-binding conduction band, which incorporates the formation of spin domains. Based on this model, predictions are made concerning the possibility of inducing magnetization of the spin lattices by illuminating the material with circularly polarized light.

MAGNETIC RESONANT X-RAY DIFFRACTION APPLIED TO THE STUDY OF EuTe FILMS AND MULTILAYERS

In this work we use magnetic resonant x-ray diffraction to study the magnetic properties of a 1.5 mu m EuTe film and an EuTe/PbTe superlattice (SL). The samples were grown by molecular beam epitaxy on (111) oriented BaF(2) substrates. The measurements were made at the Eu L(2) absorption edge, taking profit of the resonant enhancement of more than two orders in the magnetically diffracted intensity. At resonance, high counting rates above 11000 cps were obtained for the 1.5 gm EuTe film, allowing to check for the type II antiferromagnetic order of EuTe. An equal population of the three possible in-plane magnetic domains was found. The EuTe/PbTe SL magnetic peak showed a satellite structure, indicating the presence of magnetic correlations among the 5 ML (monolayers) EuTe layers across the 15 ML PbTe non-magnetic spacers. The temperature dependence of the integrated intensities of the film and the SL yielded different Neel temperatures T(N). The lower T(N) for the SL is explained considering the higher influence of the surface atoms, with partial bonds lost.

MAGNETIC FIELD INDUCED NEAR-BAND-GAP OPTICAL PROPERTIES IN EuTe LAYERS

The magnetic response of the near-band-edge optical properties is studied in EuTe layers. In several magneto-optical experiments, the absorption and emission are described as well as the related Stokes shift. Specifically, we present the first experimental report of the photoluminescence excitation (PLE) spectrum in Faraday configuration. The PLE spectra shows to be related with the absorption spectra through the observation of resonance between the excitation light and the zero-field band-gap. A new emission line appears at 1.6 eV at a moderate magnetic field in the photoluminescence (PL) spectra. Furthermore, we examine the absorption and PL red-shift induced by the magnetic field in the light of the d-f exchange interaction energy involved in these processes. Whereas the absorption red-shift shows a quadratic dependence on the field, the PL red-shift shows a linear dependence which is explained by spin relaxation of the excited state.