arXiv subjects

Sylvie Huet

Publications and source records attributed to Sylvie Huet.

22 records · Page 2Linked to original sources

arXiv2011

A commuting network model: going to the bulk

The influence of commuting in socio-economic dynamics increases constantly. Analysing and modelling the networks formed by commuters to help decision-making regarding the land-use has become crucial. This paper presents a simple spatial interaction simulated model with only one parameter. The proposed algorithm considers each individual who wants to commute, starting from their living place to all their workplaces. It decides where the location of the workplace following the classical rule inspired from the gravity law consisting in a compromise between the job offers and the distance to the jobs. The further away the job offer is, the more important it must be in order to be considered. Inversely, only the quantity of offers is important for the decision when these offers are close. The paper also presents a comparative analysis of the structure of the commuting networks of the four European regions to which we apply our model. The model is calibrated and validated on these regions. Results from the analysis shows that the model is very efficient in reproducing most of the statistical properties of the network given by the data sources.

physics.soc-ph↗

arXiv2009

An iterative approach for generating statistically realistic populations of households

Background: Many different simulation frameworks, in different topics, need to treat realistic datasets to initialize and calibrate the system. A precise reproduction of initial states is extremely important to obtain reliable forecast from the model. Methodology/Principal Findings: This paper proposes an algorithm to create an artificial population where individuals are described by their age, and are gathered in households respecting a variety of statistical constraints (distribution of household types, sizes, age of household head, difference of age between partners and among parents and children). Such a population is often the initial state of microsimulation or (agent) individual-based models. To get a realistic distribution of households is often very important, because this distribution has an impact on the demographic evolution. Usual techniques from microsimulation approach cross different sources of aggregated data for generating individuals. In our case the number of combinations of different households (types, sizes, age of participants) makes it computationally difficult to use directly such methods. Hence we developed a specific algorithm to make the problem more easily tractable. Conclusions/Significance: We generate the populations of two pilot municipalities in Auvergne region (France), to illustrate the approach. The generated populations show a good agreement with the available statistical datasets (not used for the generation) and are obtained in a reasonable computational time.

cs.MA↗

arXiv2009

Gaussian model selection with an unknown variance

Let $Y$ be a Gaussian vector whose components are independent with a common unknown variance. We consider the problem of estimating the mean $μ$ of $Y$ by model selection. More precisely, we start with a collection $\mathcal{S}=\{S_m,m\in\mathcal{M}\}$ of linear subspaces of $\mathbb{R}^n$ and associate to each of these the least-squares estimator of $μ$ on $S_m$. Then, we use a data driven penalized criterion in order to select one estimator among these. Our first objective is to analyze the performance of estimators associated to classical criteria such as FPE, AIC, BIC and AMDL. Our second objective is to propose better penalties that are versatile enough to take into account both the complexity of the collection $\mathcal{S}$ and the sample size. Then we apply those to solve various statistical problems such as variable selection, change point detections and signal estimation among others. Our results are based on a nonasymptotic risk bound with respect to the Euclidean loss for the selected estimator. Some analogous results are also established for the Kullback loss.

math.ST↗

arXiv2005

Testing convex hypotheses on the mean of a Gaussian vector. Application to testing qualitative hypotheses on a regression function

In this paper we propose a general methodology, based on multiple testing, for testing that the mean of a Gaussian vector in R^n belongs to a convex set. We show that the test achieves its nominal level, and characterize a class of vectors over which the tests achieve a prescribed power. In the functional regression model this general methodology is applied to test some qualitative hypotheses on the regression function. For example, we test that the regression function is positive, increasing, convex, or more generally, satisfies a differential inequality. Uniform separation rates over classes of smooth functions are established and a comparison with other results in the literature is provided. A simulation study evaluates some of the procedures for testing monotonicity.

math.ST↗