zbMATH — the first resource for mathematics

Estimation semi-paramétrique d’un modèle autorégressif stationnaire multiindice non nécessairement causal. (Semi-parametric estimation of a stationary, multi-indexed, non necessarily causal autoregressive process). (French) Zbl 0719.62100
The author considers a stationary autoregressive process whose index is defined on a multidimensional lattice. A procedure for estimation of the autoregressive coefficients is presented for the case when the autoregression is not necessarily causal and the noise process is not Gaussian. Asymptotic properties and robustness are discussed and an estimator of the order of such a process is presented.
62M10 Time series, auto-correlation, regression, etc. in statistics (GARCH)
62G07 Density estimation
62G20 Asymptotic properties of nonparametric inference
86A15 Seismology (including tsunami modeling), earthquakes
62M09 Non-Markovian processes: estimation
93E12 Identification in stochastic control theory
62G35 Nonparametric robustness
Full Text: Numdam EuDML