Didier A. Girard homepage

POSITION

CNRS Researcher, at LJK laboratory, D.A.T.A. Department, IPS team.

Principal Investigator for several CNRS-CEA Projects, including : Nonparametric estimation by multidimensional spline.
 

THEMES

Statistical Theory and Methodology, Computational Statistics, Inverse Problems, Inference for Stochastic Processes.
keywords: adaptive tuning; multidimensional smoothing spline; bias-variance tradeoff; generalized cross-validation; radial basis functions; numerical methods for large data sets; variational data assimilation; tomography; randomized trace; simulation based inference; measurement errors; Matern process; infill asymptotics.
 

SELECTED PUBLICATIONS

  • Optimal regularized reconstruction in computerized tomography. SIAM J. Scientific and Statistical Computing, vol. 8, 6, pp. 934-950, 1987.
  • Un algorithme rapide pour le calcul de la trace de l'inverse d'une grande matrice. Research report RR665-M, TIM3_IMAG (1987).pdf. The correlated-sampling extension suggested here has proved useful for heteroskedastic cases, as in "Letters to the Editor: Comment on O'Sullivan" JASA, 88(424), 1993, p. 1478-1479.
  • A fast 'Monte-Carlo cross-validation' procedure for large least squares problems with noisy data. RR 687-M, TIM3_IMAG (1987); and Numer. Math., vol. 56, pp. 1-23, 1989.
  • Asymptotic optimality of the fast randomized versions of GCV and CL in ridge regression and regularization. The Annals of Statistics, vol. 19, 4, pp. 1950-1963, 1991.
  • The fast Monte-Carlo cross-validation and CL procedures: comments, new results and application to image recovery problems (with discussion by seven authors and a rejoinder). Computational Statistics, vol. 10, pp. 205-258, 1995.
  • The minimum "recontruction-error" choice of regularization parameters: some more efficient methods and their application to deconvolution problems. SIAM J. Scientific and Statistical Computing, vol. 16, pp. 1387-1403, 1995. (coauthor L. Desbat)
  • Asymptotic comparison of (partial) cross-validation, GCV and randomized GCV in non-parametric regression. The Annals of Statistics, vol. 26, pp. 315-334, 1998.
  • Estimating the accuracy of (local) cross-validation via randomised GCV choices in kernel or smoothing spline regression. J. Nonparametric Statistics. vol. 22, pp. 41-64, 2010.
  • Efficiently estimating some common geostatistical models by 'energy-variance matching' or its randomized 'conditional-mean' versions. Spatial Statistics. Available online 23 March 2017. http://dx.doi.org/10.1016/j.spasta.2017.01.001.
  • Asymptotic Near-Efficiency of the 'Gibbs-Energy and Empirical-Variance' Estimating Functions for Fitting Matern Models, I: Densely sampled processes, & II: Accounting for measurement errors via conditional GE mean. Statistics and Probability Letters & https://arxiv.org/pdf/0909.1046v3.pdf.