Search

In multivariate statistics, the question of finding direct interactions can be formulated as a problem of network inference - or network reconstruction - for which the Gaussian graphical model (GGM) provides a canonical framework. Unfortunately, the Gaussian assumption does not apply to count data which are encountered in domains such as genomics, social sciences or ecology. To circumvent this limitation, state-of-the-art approaches use two-step strategies that first transform counts to pseudo Gaussian observations and then apply a (partial) correlation-based approach from the abundant literature of GGM inference. We adopt a different stance by relying on a latent model where we directly model counts by means of Poisson distributions that are conditional to latent (hidden) Gaussian correlated variables. In this multivariate Poisson lognormal-model, the dependency structure is completely captured by the latent layer. This parametric model enables to account for the effects of covariates on the counts. To perform network inference, we add a sparsity inducing constraint on the inverse covariance matrix of the latent Gaussian vector. Unlike the usual Gaussian setting, the penalized likelihood is generally not tractable, and we resort instead to a variational approach for approximate likelihood maximization. The corresponding optimization problem is solved by alternating a gradient ascent on the variational parameters and a graphical-Lasso step on the covariance matrix. We show that our approach is highly competitive with the existing methods on simulation inspired from microbiological data. We then illustrate on three various data sets how accounting for sampling efforts via offsets and integrating external covariates (which is mostly never done in the existing literature) drastically changes the topology of the inferred network.

The ability of the continuous double auction to attain the competitive equilibrium outcome is one of the most robust results in experimental economics. In this paper, we study the sensitivity of the behavior of the double auction to the imposition of a monetary cost on the submission of offers to buy and sell. We find that, even with the transaction cost, prices converge toward the competitive equilibrium level. However, the presence of the transaction cost leads to lower quantity exchanged and market efficiency than in the competitive equilibrium.

absent

Segmentation methods have been successfully applied to the mapping of chromosomal abnormalities when using CGH microarrays. Most current methods deal with one CGH profile only, and do not integrate multiple arrays, whereas the CGH microarray technology becomes widely used to characterize chromosomal defaults at the cohort level. We present CGHSeg, an R package that is devoted to the analysis of CGH profiles at the individual and at the cohort levels. This package performs segmentation in multiple CGH profiles in the framework of linear models, and multivariate segmentation/clustering for the joint characterization of aberration types (status assignment of regions based on the cohort). Overall, linear models offer a unified framework for the joint analysis of multiple CGH profiles, and we will show how they can be used to link the experience acquired in the field of expression arrays (normalization, experimental design) with array CGH data analysis.

Traduction anglaise de l'ouvrage "la génomique en biologie végétale" (2004); absent

The objective of this paper is to identify the determinants of the decision to innovate in Taiwan. Three “innovation strategies” are considered : doing R&D only, importing technology only, and combining both. We estimate a Bivariate Probit on a panel of more than 27000 Taiwanese manufacturing firms observed from 1992 to 1995. Results suggest that the decision to R&D over the period was influenced by the prior changes in exportations at the industry level, whereas the decision to import technology is affected by the current changes. We identify a non-linear relationship between firm size and innovation. Moreover, older firms tend to innovate less, whereas market structure doesn’t affect the decision to innovate. These two results change when only high-tech industries are considered : the effect of firm’s age becomes insignificant, whereas a more concentrated market structure is shown to increase the probability to innovate.

This paper investigates the relationship between firms’ innovation practices and performance in Taiwan. Using a panel of 4000 firms, we examine the effects of importing technology (versus doing R&D) on Total Factor Porductivity (TFP) growth. The relationship between these two innovation strategies is also explored. We find that R&D strongly contributes to the growth of TFP, whereas the importation of technology is only weakly significant, which makes it difficult to qualify the type of relationship (complementarity or substitutability) that exists between the two innovation strategies.

absent

A whole branch of the economic literature suggests that institutional differences between and inside educational systems may have a larger influence on students performance than the amount of resources devoted to schooling. In this paper, we use the PISA 2000 international OECD data to evaluate the impacts of organizational and institutional factors on students performance. We estimate an education production function with country fixed-effect and school random-effect. We find that, alongside individual characteristics, school autonomy in decisions regarding the recruitment of new personnel as well as pedagogical training strongly affect students performance. On the contrary, measures of school resources and standardised evaluation of students have no consistent effect.

Consumer behaviour with regard to gmo's in the food supply: an experimental study - In this paper we study the French consumer behaviors with regard to the use of genetically modified organisms in the food supply. We report the results of an economic experiment in which we elicited willingness-to-pay information for products that contained gmo's and others that were GMO-free at various thresholds. A well established European survey reveals public opinion on the same subject. Participants in both the experiment and the survey were a demographically representative sample of French urban households. We find that actual consumer behavior exhibits a lower level of hostility toward GMO's than that suggested by the survey., Dans cet article, les comportements effectifs des consommateurs concernant l'utilisation de produits génétiquement modifiés dans l'alimentation humaine sont étudiés. Il rend compte des résultats d'une expérience économique dans laquelle sont observées les propensions à payer effectives des consommateurs pour des produits alimentaires contenant des ogm et d'autres n'en contenant pas, avec différents niveaux de seuils de contamination fortuite. Un sondage européen bien établi révèle les opinions sur ces questions. Les participants à l'expérience et les sondés sont démographiquement représentatifs des chefs de ménages urbains français. Les comportements effectifs des consommateurs s'avèrent être plus nuancés et plus hétérogènes que les opinions révélées par le sondage., Noussair Charles, Robin Stéphane, Ruffieux Bernard. Comportement des consommateurs face aux aliments «avec OGM» et «sans OGM»: une étude expérimentale. In: Économie rurale. N°266, 2001. pp. 30-44.

The inefficiency of a chain of monopolies: an experimental study in a repeated-negotiation situation by Alexis Garapin, Michel Hollard, Stéphane Robin and Bernard Ruffieux Theory (bilateral monopoly, dual marginalisation) shows that a chain of monopolies is less efficient than an integrated monopoly. This result is valid both the joint profits of the bilateral monopolies and in terms of collective surplus. The goal of the present paper is to test the relevance of these theoretical conclusions in an experimental framework. The experimental results only partially confirm the theoretical conclusions. While the merger increases the welfare efficiency of the structure, it does not always increase the joint profit of the monopolies., L'inefficacité d'une chaîne de monopoles : une étude expérimentale en situation de négociation répétée par Alexis Garapin, Michel Hollard, Stéphane Robin et Bernard Ruffieux La théorie (monopole bilatéral, double marginalisation) affirme que l'efficacité d'une chaîne de monopoles est inférieure à celle d'un monopole intégré. Ce résultat est valable tant du point de vue des monopoleurs bilatéraux que du point de vue du surplus collectif. Cet article teste la validité de ces résultats théoriques en situation expérimentale. Les résultats de l'expérience ne valident que partiellement les résultats théoriques. La fusion du monopole bilatéral dans un monopole simple conduit bien à un accroissement de l'efficacité en termes de surplus collectif, mais n'entraîne pas toujours une augmentation du profit joint des deux monopoleurs., Garapin Alexis, Hollard Michel, Robin Stéphane, Ruffieux Bernard. L'inefficacité d'une chaîne de monopoles : une étude expérimentale en situation de négociation répétée. In: Économie & prévision, n°145, 2000-4. pp. 1-18.

Resume Dans cet article, nous etudions les differences entre l’opinion publique et les comportements effectifs des consommateurs concernant l’utilisation de produits genetiquement modifies dans l’alimentation humaine. Nous rendons compte des resultats d’une experience economique dans laquelle nous observons les propensions a payer effectives des consommateurs pour des produits contenant des ogm et n’en contenant pas a differents niveaux de seuils de contamination. Les opinions relatives aux memes produits ont ete etudiees par sondage. Les participants de l’experience et du sondage sont des consommateurs demographiquement representatifs des menages urbains francais. Les resultats montrent que les comportements effectifs sont plus nuances et plus heterogenes que les opinions. Nous fournissons une analyse economique explicative de ces differences et discutons leurs consequences en termes de politique d’etiquetage de ces produits.

Trust, reciprocity, and cheap talk Standard microeconomics implies a particular view of trust between agents: it can only exist under incentive mechanisms. In the experiment reported in this paper, the Dynamic Investment Game, in which such standard incentive mechanisms are not present but in which trust can lead to mutual gains, we find that certain individuals exhibit trust and thereby induce others to reciprocate. A static game theoretic model with incomplete information which contains a concept of fairness equilibrium more accurately describes behaviour in the experiment than the standard model. We further study behavior in this game with a second series of experiments in which cheap talk is permitted. Our results confirm the existence of incomplete information and reveal certain difficulties in coordinating on Pareto-optimal equilibria., Confiance, réciprocité et cheap talk La microéconomie standard donne une vision particulière des relations de confiance : elles ne peuvent exister sans mécanisme incitatif. Dans un cadre expérimental où aucun support incitatif de la théorie standard n'est présent et où la confiance permet un gain mutuel, le jeu dynamique de l'investissement montre que certains individus sont disposés à faire confiance et que d'autres sont incités à la réciprocité en présence de manifestations de confiance. Un modèle de jeu statique, reposant sur un concept d'équilibre de bienveillance réciproque, révèle une situa­tion d'information incomplète et éclaire les comportements des joueurs. Une seconde expérience introduit une communication cheap talk. Les résultats confirment l'existence d'informations incomplètes et révèlent les difficultés que rencontrent les joueurs à se coordonner sur des issues Pareto-optimales., Meidinger Claude, Robin Stéphane, Ruffieux Bernard. Confiance, réciprocité et cheap talk. In: Revue économique, volume 50, n°1, 1999. pp. 5-44.

Nous décrivons un appareil permettant d'évaporer sous ultra-vide (5 × 10—10 Torr) un matériau sous forme de couche mince et de faire des mesures optiques dans l'ultraviolet jusqu'à 1 200 Å. Les couches minces sont étudiées dans l'enceinte où elles ont été réalisées. Nous pouvons mesurer la transmission du rayonnement incident ou sa réflexion suivant deux angles d'incidence (20 degrés et 60 degrés) et en déduire les constantes optiques (n, indice de réfraction ; k, indice d'extinction) du matériau étudié [1].

International audience; Despite the vast research on the flexibility of buildings consumption, current literature is more about predicting the impacts of energy flexibility than fo-cusing on its modeling. In this paper, a methodology is provided to go from data-driven modeling of the load consumption to an optimization problem with a Mixed-Integer Linear Programming (MILP) formulation. Illustrated on an Energy Intensive Industries (EII) with an economic point of view, the methodology is suitable for any consumption site, allowing optimal energy planning studies at the district scale. Thus, it facilitates the definition of flexibility strategies to exploit the complementary of uses of the districts .