Associação Brasileira de Estatística
Repositório da Produção USP (Univ. de São Paulo)Not a member yet
45573 research outputs found
Sort by
Tubular neighborhoods and continuation of Morse decompositions
No full textWe prove a continuation result for Morse decompositions under tubular singular semiflow perturbations, which generalizes a corresponding result from Carbinatto and Rybakowski [Morse decompositions in the absence of uniqueness, II. Topol. Methods Nonlinear Anal.22 (2003), 15–51] and is applicable to cases in which the phase space of the perturbed semiflow is not necessarily homeomorphic to a product of metric spaces having as a factor the phase space of the limiting semiflow. We apply this result to singularly perturbed second-order differential equations on differential manifolds
The identification of indicators of sentiment using a multi-view self-training algorithm
No full textEste artigo apresenta um algoritmo de “multi-view self-training” , que identifica os indicadores de sentimento por: 1. extração relações causais, 2. As relações causais classificação em uma categoria sentimento, 3. agrupamento causas comuns e 4. atribuindo categorias sentimento a causas comuns para criar um distribuição sentimento para cada causa comum. Uma avaliação manual global da estratégia descobriu que ele tinha uma precisão de 70,00%.FAPESP (grant number: 11/20451-1
Stability and clustering for lattice many-body quantum hamiltonians with multiparticle potentials
No full textWe analyze a quantum system of N identical spinless particles of mass m, in the\ud
lattice Zd , given by a Hamiltonian HN = TN +VN , with kinetic energy TN ≥ 0 and potential\ud
VN = VN,2 +VN,3 composed of attractive pair and repulsive 3-body contact-potentials. This\ud
Hamiltonian is motivated by the desire to understand the stability of quantum field theories,\ud
with massive single particles and bound states in the energy-momentum spectrum, in terms\ud
of an approximate Hamiltonian for their N-particle sector. We determine the role of the\ud
potentials VN,2 and VN,3 on the physical stability of the system, such as to avoid a collapse\ud
of the N particles. Mathematically speaking, stability is associated with an N-linear lower\ud
bound for the infimum of the HN spectrum, σ(HN ) ≥ −cN, for c > 0 independent of N.\ud
For VN,3 = 0, HN is unstable, and the system collapses. If VN,3 = 0, HN is stable and,\ud
for strong enough repulsion, we obtain σ(HN ) ≥ −c \ud
N, where c \ud
N is the energy of (N/2)\ud
isolated bound pairs. This result is physically expected. A much less trivial result is that,\ud
as N varies, we show [ σ(VN )/N ] has qualitatively the same behavior as the well-known\ud
curve for minus the nuclear binding energy per nucleon. Moreover, it turns out that there\ud
exists a saturation value Ns of N at and above which the system presents a clustering: the N\ud
particles distributed in two fragments and, besides lattice translations of particle positions,\ud
there is an energy degeneracy of all two fragments with particle numbers Nr and Ns − Nr,\ud
with Nr = 1,..., Ns − 1.CNP
Evaluation of multiclass novelty detection algorithms for data streams
No full textData stream mining is an emergent research area that investigates knowledge extraction from large amounts of continuously generated data, produced by non-stationary distribution. Novelty detection, the ability to identify new or previously unknown situations, is a useful ability for learning systems, especially when dealing with data streams, where concepts may appear, disappear, or evolve over time. There are several studies currently investigating the application of novelty detection techniques in data streams. However, there is no consensus regarding how to evaluate the performance of these techniques. In this study, we propose a new evaluation methodology for multiclass novelty detection in data streams able to deal with: i) unsupervised learning, which generates novelty patterns without an association with the true classes, where one class may be composed of a novelty set, ii) confusion matrix that increases over time, iii) confusion matrix with a column representing unknown examples, i.e., those not explained by the model, and iv) representation of the evaluation measures over time. We propose a new methodology to associate the novelty patterns detected by the algorithm, in an unsupervised fashion, with the true classes. Finally, we evaluate the performance of the proposed methodology through the use of known novelty detection algorithms with artificial and real data sets.Sibila research project (NORTE-07-0124-FEDER-000059), financed by North Portugal Regional Operational Programme (ON.2 O Novo Norte), under the National Strategic Reference Framework (NSRF), through the Development Fund (ERDF)Fundação para a Ciência e a Tecnologia (FCT), and by European Commission through the project MAESTRA (Grant number ICT- 2013-612944)CAPESCNPqFAPES
Long-time behavior of a class of thermoelastic plates with nonlinear strain
No full textIn recent years a class of vibrating plates with nonlinear strain of p-Laplacian type was studied by several authors. The present paper contains a first thermoelastic model of that class of problems including both Fourier and non-Fourier heat laws. Our main result establishes the existence of global and exponential attractors for the strongly damped problem through a stabilizability inequality. In addition, for the weakly damped problem, we establish the exponential stability of its Galerkin semiflows.CNPq (Grant No. 441414/2014-1)FAPESP (Grant Nos.2012/24266-7 and 2013/07039-0)Fundação Araucária (Grant No. 308/2012)Natural Science Foundation of China (Grant No. 11271336
Evaluating word embeddings and a revised corpus for part-of-speech tagging in Portuguese
No full textBackground: Part-of-speech tagging is an important preprocessing step in many natural language processing applications. Despite much work already carried out in this field, there is still room for improvement, especially in Portuguese. We experiment here with an architecture based on neural networks and word embeddings, and that has achieved promising results in English. \ud
Methods: We tested our classifier in different corpora: a new revision of the Mac-Morpho corpus, in which we merged some tags and performed corrections and two previous versions of it. We evaluate the impact of using different types of word embeddings and explicit features as input. \ud
Results: We compare our tagger’s performance with other systems and achieve state-of-the-art results in the new corpus. We show how different methods for generating word embeddings and additional features differ in accuracy.\ud
Conclusions: The work reported here contributes with a new revision of the Mac-Morpho corpus and a state-of-the-art new tagger available for use out-of-the-box
Comparing hard and overlapping clusterings
No full textSimilarity measures for comparing clusterings is an important component, e.g., of evaluating clustering algorithms, for consensus clustering, and for clustering stability assessment. These measures have been studied for over 40 years in the domain of exclusive hard clusterings (exhaustive and mutually exclusive object sets). In the past years, the literature has proposed measures to handle more general clusterings (e.g., fuzzy/probabilistic clusterings). This paper provides an overview of these new measures and discusses their drawbacks. We ultimately develop a corrected-for-chance measure (13AGRI) capable of comparing exclusive hard, fuzzy/probabilistic, non-exclusive hard, and possibilistic clusterings. We prove that 13AGRI and the adjusted Rand index (ARI, by Hubert and Arabie) are equivalent in the exclusive hard domain. The reported experiments show that only 13AGRI could provide both a fine-grained evaluation across clusterings with different numbers of clusters and a constant evaluation between random clusterings, showing all the four desirable properties considered here. We identified a high correlation between 13AGRI applied to fuzzy clusterings and ARI applied to hard exclusive clusterings over 14 real data sets from the UCI repository, which corroborates the validity of 13AGRI fuzzy clustering evaluation. 13AGRI also showed good results as a clustering stability statistic for solutions produced by the expectation maximization algorithm for Gaussian mixture. Implementation and supplementary figures can be found at http://sn.im/25a9h8u
Differentiable positive definite kernels on two-point homogeneous spaces
No full textIn this work we study continuous kernels on compact two-point homogeneous spaces which are positive definite and zonal (isotropic). Such kernels were characterized by R. Gangolli some forty years ago and are very useful for solving scattered data interpolation problems on the spaces. In the case the space is the d-dimensional unit sphere, J. Ziegel showed in 2013 that the radial part of a continuous positive definite and zonal kernel is continuously differentiable up to order ⌊(d−1)/2⌋ in the interior of its domain. The main issue here is to obtain a similar result for all the other compact two-point homogeneous spaces.CNPq (grant 141908/2015-7)FAPESP (grant 2014/00277-5
Sandflies (Diptera, Psychodidae) from forest areas in Botucatu municipality, central western São Paulo State, Brazil
No full textAbstract
Background
The study of the distribution and ecology of sandfly species is essential for epidemiological surveillance and estimation of the transmission risk of Leishmania spp. infection.
Findings
In the present study, sandflies were captured in native fragmented forest areas in Rubião Júnior district, Botucatu municipality, São Paulo state, Brazil, between September 2001 and January 2005. A minimum of two automatic light traps were installed per night from 6 pm to 8 am, in different months, resulting in approximately 900 collecting hours. During this period, 216 sandfly specimens of sixteen species were captured. Pintomyia monticola and Brumptomyia guimaraesi were the most abundant with 56 specimens (25.93%) captured per species, followed by Pintomyia fischeri 28 (12.96%) and Psathyromyia pascalei 18 (8.33%). Other captured species were Lutzomyia amarali, Sciopemyia sordellii, Psathyromyia aragaoi, Nyssomyia whitmani, Migonemyia migonei, Pintomyia bianchigalatiae, Pintomyia misionensis, Brumptomyia carvalheiroi, Brumptomyia cardosoi, Brumptomyia cunhai, Brumptomyia nitzulescui, Brumptomyia brumpti and Brumptomyia spp. represented by 58 (26.85%) specimens.
Conclusions
Although less frequently found, the presence of Pintomyia fischeri, Nyssomyia whitmani and Migonemyia migonei, known vectors of Leishmania braziliensis, indicates risk of American cutaneous leishmaniasis occurrence. Moreover, the absence of Lutzomyia longipalpis-the main vector of Leishmania infantum chagasi, which is the agent of American visceral leishmaniasis-suggests that there is no risk of introduction and establishment of this disease in the studied area
Proteomic analysis of total cellular proteins of human neutrophils
No full textAbstract
Background
Neutrophils are the most abundant leukocytes in peripheral blood and represent one of the most important elements of innate immunity. Recent subcellular proteomic studies have focused on the identification of human neutrophil proteins in various subcellular membrane and granular fractions. Although there are relatively few studies dealing with the analysis of the total extract of human neutrophils, many biological problems such as the role of chemokines, adhesion molecules, and other activating inputs involved in neutrophil responses and signaling can be approached on the basis of the identification of the total cellular proteins.
Results
Using gel-LC-MS/MS, 251 total cellular proteins were identified from resting human neutrophils. This is more than ten times the number of proteins identified by an initial proteome analysis of human neutrophils and almost five times the number of proteins identified by the first 2-DE map of extracts of rat polymorphonuclear leukocytes. Most of the proteins identified in the present study are well-known, but some of them, such as neutrophil-secreted proteins and centaurin beta-1, a cytoplasmic protein involved in the regulation of NF-κB activity, are described here for the first-time.
Conclusion
The present report provides new information about the protein content of human neutrophils. Importantly, our study resulted in the discovery of a series of proteins not previously reported to be associated with human neutrophils. These data are relevant to the investigation of comparative pathological states and models for novel classes of pharmaceutical drugs that could be useful in the treatment of inflammatory disorders in which neutrophils participate