Search

The AOPN tomatoes and cucumbers carries the DEPHY Serre project with the objective of zero pesticides for these two major vegetables in a soilless system while maintaining economic and environmental performances. The realization of a survey at the start of the project, performed in each production area, identified the mains pests encountered for these crops. Analytic experimentations occurred in two experimental stations, aiming to identify bioproducts of interest as an alternative to chemical pesticides. In parallel, the implementation of monitoring systems was set up among 5 pilot sites (3 for tomatoes and 2 for cucumbers). This stage aimed to validate the results obtained in the experimental stations and to perform a systemic approach by integrating several levers with a global strategy of crop protection. The conducted trials in the stations identified several biocontrol products against tomatoes botrytis (fungi, bacteria or antagonist yeast), and powdery mildew for both crops (drying products which have a direct action against the pathogen or products stimulating the plant defense system). On the pilot sites, the trials enabled to tune the global protection strategies set up in tomato crops with a focus against whiteflies by combining several technics: releasing periods of Macrolophus regarding the crop calendar, interest of feeding at the start of the season, interest of reduced deleafing in release areas, use of biocontrol products, yellow sticky traps for mass trapping, aspiration around the plant's head, etc., The AOPN tomatoes and cucumbers carries the DEPHY Serre project with the objective of zero pesticides for these two major vegetables in a soilless system while maintaining economic and environmental performances. The realization of a survey at the start of the project, performed in each production area, identified the mains pests encountered for these crops. Analytic experimentations occurred in two experimental stations, aiming to identify bioproducts of interest as an alternative to chemical pesticides. In parallel, the implementation of monitoring systems was set up among 5 pilot sites (3 for tomatoes and 2 for cucumbers). This stage aimed to validate the results obtained in the experimental stations and to perform a systemic approach by integrating several levers with a global strategy of crop protection. The conducted trials in the stations identified several biocontrol products against tomatoes botrytis (fungi, bacteria or antagonist yeast), and powdery mildew for both crops (drying products which have a direct [br/] action against the pathogen or products stimulating the plant defense system). On the pilot sites, the trials enabled to tune the global protection strategies set up in tomato crops with a focus against whiteflies by combining several technics: releasing periods of Macrolophus regarding the crop calendar, interest of feeding at the start of the season, interest of reduced deleafing in release areas, use of biocontrol products, yellow sticky traps for mass trapping, aspiration around the plant's head, etc.

The estimation of optimal support boundaries under the monotonicity constraint is relatively unexplored and still in full development. This article examines a new extreme-value based model which provides a valid alternative for complete envelopment frontier models that often suffer from lack of precision, and for purely stochastic ones that are known to be sensitive to model misspecification. We provide different motivating applications including the estimation of the minimal cost in production activity and the assessment of the reliability of nuclear reactors.

In 1985 Hosking et al. estimated with the so-called Probability-Weighted Moments (PWM) method the parameters of the Generalized Extreme Value (GEV) distribution, the latter being classically fitted to maxima of sequences of independent and identically distributed random variables. Their approach is still very popular in hydrology and climatology because of its conceptual simplicity, its easy implementation and its good performance for most distributions encountered in geosciences. Its main drawback resides in its limitations when applied to strong heavy-tailed densities. Whenever the GEV shape parameter is larger than 0.5, the asymptotic properties of the PWMs cannot be derived and consequently, asymptotic confidence intervals cannot be obtained. To broaden the validity domain of the PWM approach, we take advantage of a recent extension of PWM to a larger class of moments, called Generalized PWM (GPWM). This allows us to derive the asymptotic properties of our estimators for larger values of the shape parameter. The performance of our approach is illustrated by studying simulations of small, medium and large GEV samples. Comparisons with other GEV estimation techniques used in hydrology and climatology are performed., REVSTAT-Statistical Journal, Vol. 6 No. 1 (2008): REVSTAT-Statistical Journal

The estimation of optimal support boundaries under the monotonicity constraint is relatively unexplored and still in full development. This article examines a new extreme-value based model which provides a valid alternative for completely envelopment frontier models that often super from lack of precision, and for purely stochastic ones that are known to be sensitive to model misspecification. We provide different motivating applications including the estimation of the minimal cost in production activity and the assessment of the reliability of nuclear reactors.

International audience; The analysis of heavy rainfall distributional properties is a complex object of study in hydrology and climatology, and it is essential for impact studies. In this paper, we investigate the question of how to optimize the spatial design of a network of existing weather stations. Our main criterion for such an inquiry is the capability of the network to capture the statistical properties of heavy rainfall described by the Extreme Value Theory. We combine this theory with a machine learning algorithm based on neural networks and a Query By Committee approach. Our resulting algorithm is tested on simulated data and applied to high‐quality extreme daily precipitation measurements recorded in France at 331 weather stations during the time period 1980–2010.

Les distributions des valeurs extremes generalisees (GEV) et de Pareto generalisees (GPD) sont tres utilisees dans l’analyse des valeurs extremes en hydrologie. La calibration de ces distributions au-dela d’un seuil pose cependant un certain nombre de problemes et en particulier celui du biais dans les queues de distributions. Une methode est proposee pour quantifier ce biais, methode basee sur la calibration d’une fonction a variations lentes. Tenir compte de cette fonction permet entre autre d’obtenir des estimateurs de quantiles plus precis. Par ailleurs, il est necessaire de supposer que les observations proviennent d’une seule etmeme distribution. Cette hypothese est satisfaite a condition que l’on puisse garantir une influence faible des inondations et des modifications artificielles telles que les structures regularisant le courant, les reservoirs, les digues, ... Dans le cas contraire, l’analyse devra etre effectuee sur la base de donnees censurees, en se basant uniquement sur les observations non influencees par ces divers phenomenes physiques. Les distributions des points «non inondes» et des points «inondes» seront alors toutes deux necessaires. Nous presenterons une approche permettant d’apprehender ce type de probleme et nous illustrerons son efficacite sur des donnees reelles de debits horaires pour la riviere Molenbeek a Erpe-Mere (Belgique). paper print ispartof: Revue de Statistique Appliquée pages:5-31 status: published

Several risk measures have been proposed in the literature, among them the marginal mean excess, defined as MME_p = \mathbb E[(Y^{(1)}-Q_1(1-p))_+|Y^{(2)}> Q_{2}(1-p)], provided \mathbb E|Y^{(1)}|< \infty, where (Y^{(1)}, Y^{(2)}) denotes a pair of risk factors, y_+:=\max(0,y), Q_j the quantile function of Y^{(j)}, j=1, 2, and p\in (0,1). In this paper we consider a generalization of this measure, where the random variables of main interest (Y^{(1)},Y^{(2)}) are observed together with a random covariate X \in \mathbb R^d, and where the Y^{(1)} excess is also power transformed. This leads to the concept of conditional marginal excess moment for which an estimator is proposed allowing extrapolation outside the data range. The main asymptotic properties of this estimator have been established, using empirical processes arguments combined with the multivariate extreme value theory. The finite sample behavior of the estimator is evaluated by a simulation experiment. We apply also our method on a vehicle insurance customer dataset.