Generalized spatial regression with differential regularization

ABSTRACT

We aim at analysing geostatistical and areal data observed over irregularly shaped spatial domains and having a distribution within the exponential family. We propose a generalized additive model that allows to account for spatially varying covariate information. The model is fitted by maximizing a penalized log-likelihood function, with a roughness penalty term that involves a differential quantity of the spatial field, computed over the domain of interest. Efficient estimation of the spatial field is achieved resorting to the finite element method, which provides a basis for piecewise polynomial surfaces. The proposed model is illustrated by an application to the study of criminality in the city of Portland, OR, USA.     

Nest design in a changing world: Great tit <Emphasis Type="Italic">Parus major</Emphasis> nests from a Mediterranean city environment as a case study

Investigations of urbanization effects on birds have focused mainly on breeding traits expressed after the nest-building stage (e.g. first-egg date, clutch size, breeding success, and offspring characteristics). Urban studies largely ignored how and why the aspects of nest building might be associated with the degree of urbanization. As urban environments are expected to present novel environmental changes relative to rural environments, it is important to evaluate how nest-building behavior is impacted by vegetation modifications associated with urbanization. To examine nest design in a Mediterranean city environment, we allowed urban great tits (Parus major) to breed in nest boxes in areas that differed in local vegetation cover. We found that different measures of nest size or mass were not associated with vegetation cover. In particular, nests located adjacent to streets with lower vegetation cover were not smaller or lighter than nests in parks with higher vegetation cover. Nests adjacent to streets contained more pine needles than nests in parks. In addition, in nests adjacent to streets, nests from boxes attached to pine trees contained more pine needles than nests from boxes attached to other trees. We suggest that urban-related alterations in vegetation cover do not directly impose physical limits on nest size in species that are opportunistic in the selection of nesting material. However, nest composition as reflected in the use of pine needles was clearly affected by habitat type and the planted tree species present, which implies that rapid habitat change impacts nest composition. We do not exclude that urbanization might impact other aspects of nest building behaviour not covered in our study (e.g. costs of searching for nest material), and that the strengths of the associations between urbanization and nest structures might differ among study populations or species.     

The effects of a change in the point system on immigration: evidence from the 2001 Quebec reform

In 2001, Quebec changed its point system, a system that selects immigrants based on specific observable characteristics. The explicit objective was to increase the number of French-speaking immigrants, with no deterioration in overall labor market performance. To achieve this, points for French and education (specifically bachelor’s degrees) were increased. In parallel, points for a subjective assessment of “adaptability” were decreased. In line with the initial objective, we find more French-speaking immigrants with bachelor’s degrees, and no worsening in labor market outcomes after the reform. These results hold in a difference-in-differences and triple differences analysis. This paper shows how point systems can be used to shape the immigrant workforce according to policy goals.     

Collaborative practices in new product development projects involving suppliers

Abstract

Collaboration with suppliers in New Product Development (NPD) is essential for business competitiveness. However, the management of such collaborative project needs to focus on interactions between team members from different companies since they performed inter-organizational activities. While the impact of collaboration with suppliers on NPD project performance has been widely studied, the investigation of practices to support daily collaborative activities between both teams has received limited attention. This paper aims at developing a framework of operational practices for successful inter-organizational collaboration in NPD. The practices enacted by both project teams in six case studies were examined according to the stages of the collaboration and a knowing perspective. The results revealed some peculiarities linked to the inter-organizational context. Practices interact with each other across stages with a specific dynamic. More particularly, early in the project, practices to justify the collaboration with the suppliers are determinant for practices related to the creation of social interactions between project members and project commitment. The results also show that boundary objects are useful formal socialization mechanisms for supporting the undertaking of collaborative work.     

Improving prediction performance of stellar parameters using functional models

This paper investigates the problem of prediction of stellar parameters, based on the star's electromagnetic spectrum. The knowledge of these parameters permits to infer on the evolutionary state of the star. From a statistical point of view, the spectra of different stars can be represented as functional data. Therefore, a two-step procedure decomposing the spectra in a functional basis combined with a regression method of prediction is proposed. We also use a bootstrap methodology to build prediction intervals for the stellar parameters. A practical application is also provided to illustrate the numerical performance of our approach.     

Tentative evaluation of the impact of public transfers on the dynamics of poverty: The case of Russia

The aim of this article is to offer a dynamic impact analysis of the system of transfers in Russia, based on a comparison of indicators of well‐being measured before and after state intervention. We shall begin by assessing the impact of public transfers on different forms of poverty and demonstrate that, while the system is seeing a fall in chronic and transitional poverty, there is very little movement between categories. We shall then evaluate the capacity of the system to keep non‐poor households from falling into poverty (protection) and to help poor households escape poverty (promotion). Several studies suggest that the Russian system of transfers is well suited to protection but has proved incapable of attaining the goal of promotion. In other words, in its current form it cannot claim to be an effective tool to combat long‐term poverty.     

Propensity score applied to survival data analysis through proportional hazards models: a Monte Carlo study

Propensity score methods are increasingly used in medical literature to estimate treatment effect using data from observational studies. Despite many papers on propensity score analysis, few have focused on the analysis of survival data. Even within the framework of the popular proportional hazard model, the choice among marginal, stratified or adjusted models remains unclear. A Monte Carlo simulation study was used to compare the performance of several survival models to estimate both marginal and conditional treatment effects. The impact of accounting or not for pairing when analysing propensity‐score‐matched survival data was assessed. In addition, the influence of unmeasured confounders was investigated. After matching on the propensity score, both marginal and conditional treatment effects could be reliably estimated. Ignoring the paired structure of the data led to an increased test size due to an overestimated variance of the treatment effect. Among the various survival models considered, stratified models systematically showed poorer performance. Omitting a covariate in the propensity score model led to a biased estimation of treatment effect, but replacement of the unmeasured confounder by a correlated one allowed a marked decrease in this bias. Our study showed that propensity scores applied to survival data can lead to unbiased estimation of both marginal and conditional treatment effect, when marginal and adjusted Cox models are used. In all cases, it is necessary to account for pairing when analysing propensity‐score‐matched data, using a robust estimator of the variance. Copyright © 2012 John Wiley & Sons, Ltd.     

Inferring large graphs using $$\ell _1$$-penalized likelihood

We address the issue of recovering the structure of large sparse directed acyclic graphs from noisy observations of the system. We propose a novel procedure based on a specific formulation of the \(\ell _1\)-norm regularized maximum likelihood, which decomposes the graph estimation into two optimization sub-problems: topological structure and node order learning. We provide convergence inequalities for the graph estimator, as well as an algorithm to solve the induced optimization problem, in the form of a convex program embedded in a genetic algorithm. We apply our method to various data sets (including data from the DREAM4 challenge) and show that it compares favorably to state-of-the-art methods. This algorithm is available on CRAN as the R package GADAG.