Analyse de sensibilité de l’effet d’un programme de prévention avec randomisation : application de trois techniques d’appariement pour balancer les groupes contrôle et expérimental : distance de Mahanalobis, score de propension et algorithme génétique

Les analyses effectuées dans le cadre de ce mémoire ont été réalisées à l'aide du module MatchIt disponible sous l’environnent d'analyse statistique R. / Statistical analyzes of this thesis were performed using the MatchIt package available in the statistical analysis environment R.

Modified Large Margin Nearest Neighbor Metric Learning for Regression

The main objective of this letter is to formulate a new approach of learning a Mahalanobis distance metric for nearest neighbor regression from a training sample set. We propose a modified version of the large margin nearest neighbor metric learning method to deal with regression problems. As an application, the prediction of post-operative trunk 3-D shapes in scoliosis surgery using nearest neighbor regression is described. Accuracy of the proposed method is quantitatively evaluated through experiments on real medical data.

El programa Proniño en Colombia: su efecto en la reducción del trabajo infantil

El objetivo de este documento es evaluar empíricamente el efecto que tiene una mayor o menor exposición del programa Proniño de la Fundación Telefónica en la cantidad de horas trabajadas de los niños, niñas y adolescentes. Tomando información desde el 2010 y hasta el 2012 se evalúa empíricamente el impacto en la duración a la exposición del tratamiento, eliminando el sesgo de selección por duración a través de la metodología de Propensity Score Matching. Los resultados muestran que la exposición al tratamiento sí logra reducir el número de horas trabajadas de los niños, niñas y adolescentes, alcanzando los niveles más altos de reducción cuando el tiempo de exposición al programa es más amplio; en este caso, tres años y particularmente para el grupo de edad comprendido entre los 12 a 14 años. Finalmente se evidencia que el programa es más efectivo en la reducción del número de horas trabajadas a la semana de los niños (hombres) que de las niñas (mujeres).

Impacto del Internet en el ingreso laboral: análisis del lugar y manejo del Internet

Este estudio analiza el efecto del uso de Internet sobre el ingreso laboral, eliminado el problema de endogeneidad que existe por motivos de selección. También se analizan los efectos seg un el lugar de acceso y las actividades realizadas en Internet. Para esto se utiliza Propensity Score Matching, tomando información para Colombia entre el año 2009 a 2011. Los resultados muestran que existe un efecto positivo, significativo y heterogéneo entre trabajadores asalariados y cuenta propia, siendo los últimos los más beneficiados. Con respecto al lugar de uso se encuentra un impacto positivo del uso del Internet en el trabajo y/o en el hogar, con una mayor magnitud en el primer sitio de acceso. Para el tema del manejo del Internet, no se encuentra efecto cuando se utiliza en actividades de entretenimiento, pero si cuando se usa en tareas educativas diariamente. Esto siguiere que usar este servicio aumenta la productividad y el capital humano, lo que se transforma en un incremento en los ingresos. Por tanto, se encuentra que el lugar de uso es una señal informativa para el mercado laboral, mientras que el impacto del manejo de esta herramienta depende de la frecuencia con que se realicen las actividades.

School management and sexual behavior of teenagers

This paper asks whether school based management may help reducing risky sexual behavior of teenagers. For this purpose we use student level data from Bogot´a to identify students from Concession School (CS), who are enrolled in public education system with a more school management autonomy at school level, and to compare them with those students at the traditional public education system. We use propensity score matching methods to have a comparable sample between pupils at CS and traditional schools. Our results show that on average the behavior of students from CS do not have a sexual behavior that differs from those in traditional public schools except for boys in CS who have a lower probability of being sexual active. However, there are important differences when heterogeneity is considered. For example we find that CS where girls per boys ratio is higher have lower teenage pregnancy rates than public schools with also high girls per boys ratios. We also find that teachers’ human capital, teacher-pupil ratio or whether school offers sexual education are also related to statistically significant differences between CS and traditional public schools.

Rendimiento acad??mico y diversidad cultural : el eje ling????stico

Resumen basado en el de la publicaci??n

El liderazgo educativo en Europa : una aproximaci??n transcultural

Resumen basado en el de la publicaci??n

Cidades sustentáveis – o que o poder local e uma formação adequada em urbanismo podem fazer?

Os sistemas de ordenamento territorial em grande medida definem aonde e quanta actividade deve ser implantada na cidade (4). Em muitas cidades as actividades se desenvolvem e localizam sem uma decisão estratégica a priori que considere o transporte , as distâncias e o consumo de energia e dos recursos naturais locais , resultando em muitos casos em padrões desequilibrados de uso e ocupação do solo. Essa decisão implica no desenvolvimento de uma gestão correcta do uso do solo urbano que suporte padrões de mobilidade urbana , associados a políticas adequadas de acessibilidades e transportes , de modo a evitar os efeitos negativos do tráfego e da poluição sobre o ambiente urbano

Application of auto regressive models of wavelet sub-bands for classifying Terahertz pulse measurements

This paper presents an approach for automatic classification of pulsed Terahertz (THz), or T-ray, signals highlighting their potential in biomedical, pharmaceutical and security applications. T-ray classification systems supply a wealth of information about test samples and make possible the discrimination of heterogeneous layers within an object. In this paper, a novel technique involving the use of Auto Regressive (AR) and Auto Regressive Moving Average (ARMA) models on the wavelet transforms of measured T-ray pulse data is presented. Two example applications are examined - the classi. cation of normal human bone (NHB) osteoblasts against human osteosarcoma (HOS) cells and the identification of six different powder samples. A variety of model types and orders are used to generate descriptive features for subsequent classification. Wavelet-based de-noising with soft threshold shrinkage is applied to the measured T-ray signals prior to modeling. For classi. cation, a simple Mahalanobis distance classi. er is used. After feature extraction, classi. cation accuracy for cancerous and normal cell types is 93%, whereas for powders, it is 98%.

Paleoamerican Morphology in the Context of European and East Asian Late Pleistocene Variation: Implications for Human Dispersion Into the New World

Early American crania show a different morphological pattern from the one shared by late Native Americans. Although the origin of the diachronic morphological diversity seen on the continents is still debated, the distinct morphology of early Americans is well documented and widely dispersed. This morphology has been described extensively for South America, where larger samples are available. Here we test the hypotheses that the morphology of Early Americans results from retention of the morphological pattern of Late Pleistocene modern humans and that the occupation of the New World precedes the morphological differentiation that gave rise to recent Eurasian and American morphology. We compare Early American samples with European Upper Paleolithic skulls, the East Asian Zhoukoudian Upper Cave specimens and a series of 20 modern human reference crania. Canonical Analysis and Minimum Spanning Tree were used to assess the morphological affinities among the series, while Mantel and Dow-Cheverud tests based on Mahalanobis Squared Distances were used to test different evolutionary scenarios. Our results show strong morphological affinities among the early series irrespective of geographical origin, which together with the matrix analyses results favor the scenario of a late morphological differentiation of modern humans. We conclude that the geographic differentiation of modern human morphology is a late phenomenon that occurred after the initial settlement of the Americas. Am J Phys Anthropol 144:442-453, 2011. (c) 2010 Wiley-Liss, Inc.

POSTMARITAL RESIDENCE PRACTICE IN SOUTHERN BRAZILIAN COASTAL GROUPS: CONTINUITY AND CHANGE

The coastal plains of the States of Parana and Santa Catarina, in Southern Brazil, were first settled around 6000 B.P. by shellmound builders, a successful fisher-hunter-gatherer population that inhabited the coastal lowlands practically unchanged for almost five thousand years. Shellmounds were typically occupied as residential sites as well as cemeteries, and are usually associated with rich alimentary zones. Around 1200 B.P., the first evidence of ceramics brought from the interior is found in coastal areas, and together with ceramics there is a progressive abandonment of shellmound construction in favor of flat cold shallow sites. Here we consider if these changes were reflected in the postmarital residence practice of coastal groups, i.e., if the arrival or intensification of contact with groups from the interior resulted in changes in this aspect of social structure among the coastal groups. To test the postmarital residence practice we analyzed within-group variability ratios between males and females, following previous studies on the topic. and between-group, correlations between Mahalanobis distances and geographic distances. The results suggest that in the pre-ceramic series a matrilocal, postmarital residential system predominated, while in the ceramic period there was a shift toward patrilocality. This favors the hypothesis that the changes experienced by coastal groups after 1200 B.P. affected not only their economy and material culture, but important aspects of their sociopolitical organization as well.

Local influence for Student-t partially linear models

In this paper we extend partial linear models with normal errors to Student-t errors Penalized likelihood equations are applied to derive the maximum likelihood estimates which appear to be robust against outlying observations in the sense of the Mahalanobis distance In order to study the sensitivity of the penalized estimates under some usual perturbation schemes in the model or data the local influence curvatures are derived and some diagnostic graphics are proposed A motivating example preliminary analyzed under normal errors is reanalyzed under Student-t errors The local influence approach is used to compare the sensitivity of the model estimates (C) 2010 Elsevier B V All rights reserved

Multivariate measurement error models based on scale mixtures of the skew-normal distribution

Scale mixtures of the skew-normal (SMSN) distribution is a class of asymmetric thick-tailed distributions that includes the skew-normal (SN) distribution as a special case. The main advantage of these classes of distributions is that they are easy to simulate and have a nice hierarchical representation facilitating easy implementation of the expectation-maximization algorithm for the maximum-likelihood estimation. In this paper, we assume an SMSN distribution for the unobserved value of the covariates and a symmetric scale mixtures of the normal distribution for the error term of the model. This provides a robust alternative to parameter estimation in multivariate measurement error models. Specific distributions examined include univariate and multivariate versions of the SN, skew-t, skew-slash and skew-contaminated normal distributions. The results and methods are applied to a real data set.

Correlations between chemical composition and provenance of Justino site ceramics by INAA

Instrumental neutron activation analysis (INAA), have been used for the definition of compositional groups of potteries from Justino site, Brazil, according to the chemical similarities of ceramic paste. The outliers were identified by means of robust Mahalanobis distance. The temper effect in the ceramic paste was studied by means of modified Mahalanobis filter. The results were interpreted by means of cluster, principal components, and discriminant analyses. This work provides contributions for the reconstruction of the prehistory of baixo Sao Francisco region, and for the reconstitution of the Brazilian Northeast ceramist population of general frame.

O problema do pixel mistura : um estudo comparativo

Os satélites para sensoriamento remoto atualmente dispoívies à comunidade científica possuem diferenies resoluções espaciais, por exemplo: SPOT 20 e 10 metros, LANDSAT-TM 30 metros e NOA-AVHRR 1100 metros. Essa resolução frequentemente não é grande o suficiente para um grande número de aplicações que necessitam de uma percepção da cena mais detalhada. Muitas vezes, no interior de uma célula de resolução (pixel) mais de uma classe ocorre. Este caso é conhecido como pixel mistura. Na classificação de imagens obtidas por sensoriamento remoto é comum a utilização de metodologias que atribuem somente uma classe a um pixel, como o procedimento clássico da máxima verossimilhança. Esse procedimento resulta frequentemente em uma estimação errônea das áreas ocupadas pelas classes presentes na cena. Em alguns casos, especialmente quando não há uma classe dominante, isto pode ser a fonte de um erro significativo. Desde o início dos anos 70, diferentes metodologias têm sido propostas para o trabalho num nível de subpixel. A grande vantagem do trabalho nesse nível é que um pixel não é necessariamente atribuído a somente uma classe. O pixel tem um grau que o correlaciona a cada classe: de zero(se a classe não ocorre no pixel) até 1 (a classe ocorre no pixel inteiro). Assim, cada pixel tem um vetor associado que estima a proporção de cada classe nele. A metodologia mais comumente utilizada considera a refletância do pixel mistura como uma combinação linear da refletância média de cada classe componente. De acordo com essa visão as refletâncias associadas às classes componentes são consideradas constantes conhecidas i.e., não são variáveis aleatórias. Assim, a proporção de cada classe no pixel é obtida pela resolução de um sistema de equações lineares. Uma outra metodologia é assumir as refletâncias que caracterizam as classes como sendo variáveis aleatórias. Nesta visão, as informações a respeito das distribuições das classes é utilizada. A estimativa das proporções de cada classe é obtida pelo vetor de proporções que maximiza a função de verossimilhança. Mais recentemente, uma visão diferente foi proposta: a utilização da lógica fuzzy. Esta metodologia utiliza o conceito de função de pertinência que é essencial à teoria dos conjuntos fuzzy. Esta função utiliza elementos com natureza estatística ou não para a estimação das proporções. No presente trabalho, duas funções de pertinência foram definidas: a primeira baseada na função densidade probabilidade gaussiana e a segunda baseada diretamente na distância de Mahalanobis. O objetivo deste estudo é avaliar cada uma das metodologias anteriores em termos de acurácia, performance e dados necessários. Para este objetivo, as metodologias foram implementadas computacionalmente e alimentadas com imagens LANDSAT-TM. Para a avaliação da acurácia dos modelos um estudo qualitativo foi executado.