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.

Poly-bagging predictors for classification modelling for credit scoring

Credit scoring modelling comprises one of the leading formal tools for supporting the granting of credit. Its core objective consists of the generation of a score by means of which potential clients can be listed in the order of the probability of default. A critical factor is whether a credit scoring model is accurate enough in order to provide correct classification of the client as a good or bad payer. In this context the concept of bootstraping aggregating (bagging) arises. The basic idea is to generate multiple classifiers by obtaining the predicted values from the fitted models to several replicated datasets and then combining them into a single predictive classification in order to improve the classification accuracy. In this paper we propose a new bagging-type variant procedure, which we call poly-bagging, consisting of combining predictors over a succession of resamplings. The study is derived by credit scoring modelling. The proposed poly-bagging procedure was applied to some different artificial datasets and to a real granting of credit dataset up to three successions of resamplings. We observed better classification accuracy for the two-bagged and the three-bagged models for all considered setups. These results lead to a strong indication that the poly-bagging approach may promote improvement on the modelling performance measures, while keeping a flexible and straightforward bagging-type structure easy to implement. (C) 2011 Elsevier Ltd. All rights reserved.

A Road Following Approach Using Artificial Neural Networks Combinations

Navigation is a broad topic that has been receiving considerable attention from the mobile robotic community over the years. In order to execute autonomous driving in outdoor urban environments it is necessary to identify parts of the terrain that can be traversed and parts that should be avoided. This paper describes an analyses of terrain identification based on different visual information using a MLP artificial neural network and combining responses of many classifiers. Experimental tests using a vehicle and a video camera have been conducted in real scenarios to evaluate the proposed approach.

Combining structure and dynamics: non-denaturing high-pressure effect on lysozyme in solution

Small-angle X-ray scattering (SAXS) and elastic and quasi-elastic neutron scattering techniques were used to investigate the high-pressure-induced changes on interactions, the low-resolution structure and the dynamics of lysozyme in solution. SAXS data, analysed using a global-fit procedure based on a new approach for hydrated protein form factor description, indicate that lysozyme completely maintains its globular structure up to 1500 bar, but significant modi. cations in the protein-protein interaction potential occur at approximately 600-1000 bar. Moreover, the mass density of the protein hydration water shows a clear discontinuity within this pressure range. Neutron scattering experiments indicate that the global and the local lysozyme dynamics change at a similar threshold pressure. A clear evolution of the internal protein dynamics from diffusing to more localized motions has also been probed. Protein structure and dynamics results have then been discussed in the context of protein-water interface and hydration water dynamics. According to SAXS results, the new configuration of water in the first hydration layer induced by pressure is suggested to be at the origin of the observed local mobility changes.

Geochronological data from the Faxinal coal succession, southern Parana Basin, Brazil: A preliminary approach combining radiometric U-Pb dating and palynostratigraphy

A radiometric zircon age of 285.4 +/- 8.6 Ma (IDTIMS U-Pb) is reported from a tonstein layer interbedded with coal seams in the Faxinal coalfield, Rio Grande do Sul, Brazil. Calibration of palynostratigraphic data with the absolute age shows that the coal depositional interval in the southern Parana Basin is constrained to the Sakmarian. Consequently, the basal Gondwana sequence in the southern part of the basin should lie at the Carboniferous-Permian boundary, not within the Sakmarian as previously considered. The new results are significant for correlations between the Parana Basin and the Argentinian Paganzo Basin (302 +/- 6 Ma and 288 +/- 7 Ma) and with the Karoo Basin, specifically with the top of the Dwyka Tillite (302 +/- 3 Ma and 299.2 +/- 3.2 Ma) and the lowermost Ecca Group (288 +/- 3 Ma and 289.6 +/- 3.8 Ma). The evidence signifies widespread latest Carboniferous volcanic activity in western Gondwana. (C) 2007 Elsevier Ltd. All rights reserved.

The interaction between atoms of Au and Cu with clean Si(111) surface: A study combining synchrotron radiation grazing incidence X-ray fluorescence analysis and theoretical calculations

In order to evaluate the interactions between Au/Cu atoms and clean Si(l 11) surface, we used synchrotron radiation grazing incidence X-ray fluorescence analysis and theoretical calculations. Optimized geometries and energies on different adsorption sites indicate that the binding energies at different adsorption sites are high, suggesting a strong interaction between metal atom and silicon surface. The Au atom showed higher interaction than Cu atom. The theoretical and experimental data showed good agreement. Crown Copyright (C) 2009 Published by Elsevier B.V. All rights reserved.

Evaluation of the maxillary premolar roots dissociation using radiographic holders with conventional and digital radiography

This in vivo study evaluated the dissociation quality of maxillary premolar roots combining variations of vertical and horizontal angulations by using X-ray holders (Rinn -XCP), and made a comparison between two types of intraoral radiography systems - conventional film (Kodak Insight, Rochester, USA) and digital radiography (Kodak RVG 6100, Kodak, Rochester, USA). The study sample was comprised of 20 patients with a total of 20 maxillary premolars that were radiographed, using the paralleling angle technique (GP), with a 20º variation of the horizontal angle (GM) and 25º variation of the horizontal angle combined with 15º vertical angle (GMV). Each image was independently analyzed by two experienced examiners. These examiners assigned a score to the diagnostic capability of root dissociation and the measurement of the distance between the apexes. Statistical data was derived using the Wilcoxon Signed Rank test, Friedman and T test. The means of the measured distances between buccal and lingual root apexes were greater for the GMV, which ranged from 2.3 mm to 3.3 mm. A statistically significant difference was found between GM and GMV when compared to GP with p < 0.01. An established best diagnostic dissociation roots image was found in the GMV. These results support the use of the anterior X-ray holders which offer a better combined deviation (GMV) to dissociate maxillary premolar roots in both radiography systems.

Accuracy of three electronic apex locators in the presence of different irrigating solutions

The present study compared the accuracy of three electronic apex locators (EALs) - Elements Diagnostic®, Root ZX® and Apex DSP® - in the presence of different irrigating solutions (0.9% saline solution and 1% sodium hypochlorite). The electronic measurements were carried out by three examiners, using twenty extracted human permanent maxillary central incisors. A size 10 K file was introduced into the root canals until reaching the 0.0 mark, and was subsequently retracted to the 1.0 mark. The gold standard (GS) measurement was obtained by combining visual and radiographic methods, and was set 1 mm short of the apical foramen. Electronic length values closer to the GS (± 0.5 mm) were considered as accurate measures. Intraclass correlation coefficients (ICCs) were used to verify inter-examiner agreement. The comparison among the EALs was performed using the McNemar and Kruskal-Wallis tests (p < 0.05). The ICCs were generally high, ranging from 0.8859 to 0.9657. Similar results were observed for the percentage of electronic measurements closer to the GS obtained with the Elements Diagnostic® and the Root ZX® EALs (p > 0.05), independent of the irrigating solutions used. The measurements taken with these two EALs were more accurate than those taken with Apex DSP®, regardless of the irrigating solution used (p < 0.05). It was concluded that Elements Diagnostic® and Root ZX® apex locators are able to locate the cementum-dentine junction more precisely than Apex DSP®. The presence of irrigating solutions does not interfere with the performance of the EALs.

Combination effect of fluoride dentifrices and varnish on deciduous enamel demineralization

The aim of this study was to evaluate the anticaries potential of 500 or 1100 ppm F dentifrices combined with fluoride varnish using a pH-cycling regimen. Seventy primary canines were covered with nail polish, leaving a 4×4 mm window on their buccal surface, and randomly assigned into 7 groups (n = 10): S: sound enamel not submitted to the pH-cycling regimen or treatment; N: negative control, submitted to the pH-cycling regimen without any treatment; D1 and D2: subjected to the pH-cycling regimen and treated twice daily with 1100 or 500 ppm F dentifrice, respectively; VF: fluoride varnish (subjected to F-varnish before and in the middle of the pH-cycling regimen); and VF+D1 and VF+D2. After 10 days, the teeth were sectioned, and enamel demineralization was assessed by cross-sectional hardness at different distances from the dental surface. Data were analyzed using a two-way ANOVA followed by Tukey's test. Dentifrice with 1100 ppm F and the combination of F-varnish with the dentifrices significantly reduced enamel demineralization compared with the negative control (p < 0.05), but the isolated effects of F-varnish and dentifrice with low concentration were not significant (p > 0.05). The effect of combining F-varnish with the dentifrices was not greater than the effect of the dentifrices alone (p < 0.05). The data suggest that the combination of F-varnish with dentifrices containing 500 and 1100 ppm F is not more effective in reducing demineralization in primary teeth than the isolated effect of dentifrice containing 1100 ppm F.

Répteis da Estação Ecológica Serra Geral do Tocantins, Brasil Central

Os Cerrados sul-americanos abrigam alta diversidade de répteis, incluindo elevado número de endemismos. No entanto, o conhecimento desta diversidade é ainda incompleto frente à acelerada transformação das paisagens naturais no Brasil central. Constituem, portanto, uma das regiões prioritárias para estudo e conservação da biodiversidade mundial. Estudos intensivos sobre a fauna de répteis do Cerrado são necessários e urgentes para melhor compreensão dos processos que levaram à sua origem e distribuição e para subsidiar ações de conservação. Por meio de métodos padronizados, amostramos duas regiões ainda inexploradas da Estação Ecológica Serra Geral do Tocantins, situada na região do Jalapão. Registramos 45 espécies de répteis para a EESGT e entorno, o que representa uma riqueza alta e comparável à de outras regiões bem amostradas do Cerrado. Curvas de acumulação e estimadores indicam que a riqueza local de lagartos e anfisbenídeos aproxima-se da riqueza real enquanto a de serpentes é subestimada. A distribuição não-aleatória das espécies na paisagem concorda com evidências anteriores sugerindo utilização diferencial dos hábitats pelos répteis. Reunindo os resultados do presente estudo com os de levantamentos prévios realizados na região, registramos 88 espécies de répteis para o Jalapão sendo oito registros novos que incluem Bachia oxyrhina uma espécie recém descrita da região. As espécies da área apresentam três padrões gerais de distribuição: (1) espécies endêmicas do Cerrado, (2) espécies compartilhadas com domínios da diagonal de formações abertas sul-americanas, e (3) espécies de ampla ocorrência, compartilhadas também com ecossistemas florestais. Prevalecem espécies de ampla distribuição, porém é grande o número de espécies típicas do Cerrado, incluindo cinco possivelmente endêmicas do Jalapão, e há contribuição importante da fauna da Caatinga. A distribuição dos répteis em escala local e regional demonstra a necessidade de considerar a heterogeneidade paisagística para o planejamento de diretrizes visando à conservação em regiões do Cerrado. Por sua grande extensão, posição biogeográfica e complexidade de relevo e tipos de hábitat, a EESGT tem papel fundamental para a preservação e conhecimento da diversidade de répteis do Cerrado.

Pre-processing for noise detection in gene expression classification data

Due to the imprecise nature of biological experiments, biological data is often characterized by the presence of redundant and noisy data. This may be due to errors that occurred during data collection, such as contaminations in laboratorial samples. It is the case of gene expression data, where the equipments and tools currently used frequently produce noisy biological data. Machine Learning algorithms have been successfully used in gene expression data analysis. Although many Machine Learning algorithms can deal with noise, detecting and removing noisy instances from the training data set can help the induction of the target hypothesis. This paper evaluates the use of distance-based pre-processing techniques for noise detection in gene expression data classification problems. This evaluation analyzes the effectiveness of the techniques investigated in removing noisy data, measured by the accuracy obtained by different Machine Learning classifiers over the pre-processed data.

Late Quaternary drainage dynamics in northern Brazil based on the study of a large paleochannel from southwestern Marajó Island

Marajó Island shows an abundance of paleochannels easily mapped in its eastern portion, where vegetation consists mostly of savannas. SRTM data make possible to recognize paleochannels also in western Marajó, even considering the dense forest cover. A well preserved paleodrainage network from the adjacency of the town of Breves (southwestern Marajó Island) was investigated in this work combining remote sensing and sedimentological studies. The palimpsest drainage system consists of a large meander connected to narrower tributaries. Sedimentological studies revealed mostly sharp-based, fining upward sands for the channelized features, and interbedded muds and sands for floodplain areas. The sedimentary structures and facies successions are in perfect agreement with deposition in channelized and floodplain environments, as suggested by remote sensing mapping. The present study shows that this paleodrainage was abandoned during Late Pleistocene, slightly earlier than the Holocene paleochannel systems from the east part of the island. Integration of previous studies with the data available herein supports a tectonic origin, related to the opening of the Pará River along fault lineaments. This would explain the disappearance of large, north to northeastward migrating channel systems in southwestern Marajó Island, which were replaced by the much narrower, south to southeastward flowing modern channels.

Preparação de membranas de acetato de celulose organomodificadas para adsorção dos íons Cu(II), Cd(II), Mn(II) e Ni(II)

Cellulose acetate polymeric membranes had been prepared by a procedure of two steps, combining the method of phase inversion and the technique of hydrolysis-deposition. The first step was the preparation of the membrane, and together was organomodified with tetraethylortosilicate and 3-aminopropyltrietoxysilane. Parameters that exert influence in the complexation of the metallic ion, as pH, time of complexation, metal concentration, had been studied in laboratory using tests of metal removal. The membranes had presented resistance mechanics and reactivity to cations, being able to be an alternative for the removal, daily pay-concentration or in the study of the lability of metals complexed.

Comércio varejista de alimentos no município de São Paulo: a evolução da legislação sanitária

O código Sanitário Municipal de Alimentos, Decreto Municipal nº 25.544, de 14 de março de 1988, regulamentou a fiscalização sanitária de g~eneros alimentícios no varejo até 26 de novembro de 2002, data em que entrou em vigor a Lei Estadual nC 10.083, de 23 de setembro de 1998, Código Sanitário do estado de São Paulo, utilizado pelos serviços municipais de vigilância, em caráter temporário, até a promulgação do atual Código Sanitário do Município de São Paulo nº 13.735, em 9 de janeiro de 2004. Este código regulamenta todos os serviços e produtos de interesse da saúde, inclusive a produção e distribuição de alimentos e água para consumo humano. Por meio do estudo dos diferentes códigos vigentes no município e de legislações esparsas, foi possível identificar as nmudanças ocorridas na legislação e nos procedimentos administrativos da vigilância sanitária do varejo de alimentos no Município de São Paulo. Concluiu-se que o Código Sanitário do Município de São Paulo é uma legislação completa e atualizada, com previsão legal de utilização de regulamentos técnicos modernos de forma combinada, especialmente os que tratam da produção e distribuição de alimentos, com ênfase nas Boas Práticas de Fabricação (BPFs)

Mobility and social network effects on extremist opinions

Understanding the emergence of extreme opinions and in what kind of environment they might become less extreme is a central theme in our modern globalized society. A model combining continuous opinions and observed discrete actions (CODA) capable of addressing the important issue of measuring how extreme opinions might be has been recently proposed. In this paper I show extreme opinions to arise in a ubiquitous manner in the CODA model for a multitude of social network structures. Depending on network details reducing extremism seems to be possible. However, a large number of agents with extreme opinions is always observed. A significant decrease in the number of extremists can be observed by allowing agents to change their positions in the network.