LAN property for some fractional type Brownian motion

Serge Cohen,Fabrice Gamboa,Céline Lacaux,Jean-Michel Loubes
DOI: https://doi.org/10.48550/arXiv.1111.1077
2011-11-04
Statistics Theory
Abstract:We study asymptotic expansion of the likelihood of a certain class of Gaussian processes characterized by their spectral density $f_\theta$. We consider the case where $f_\theta\PAR{x} \sim_{x\to 0} \ABS{x}^{-\al(\theta)}L_\theta(x)$ with $L_\theta$ a slowly varying function and $\al\PAR{\theta}\in (-\infty,1)$. We prove LAN property for these models which include in particular fractional Brownian motion %$B^\alpha_t,\: \alpha \geq 1/2$ or ARFIMA processes.
What problem does this paper attempt to address?