Abstract:We study the problem of parametric estimation for continuously observed stochastic differential equation driven by fractional Brownian motion. Under some assumptions on drift and diffusion coefficients, we construct maximum likelihood estimator and establish its the asymptotic normality and moment convergence of the drift parameter when a small dispersion coefficient vanishes.
What problem does this paper attempt to address?
The paper is primarily dedicated to addressing the following issues:
1. **Parameter Estimation Problem**: The study focuses on the parameter estimation problem for stochastic differential equations (SDEs) driven by fractional Brownian motion in the small noise asymptotic case. Specifically, for the stochastic differential equation of the form:
\[
X^{\varepsilon}_t = X_0 + \int_0^t b(X^{\varepsilon}_s, \theta_0) ds + \varepsilon \int_0^t \sigma(X^{\varepsilon}_s) dB_s, \quad t \in [0, T],
\]
where $B_t$ is a fractional Brownian motion with Hurst index $H \in (1/3, 1/2)$. This paper aims to construct a maximum likelihood type estimator and derive its asymptotic properties.
2. **Fundamental Issues in Statistical Inference**: In the small noise limit, i.e., as $\varepsilon \to 0$, the paper constructs and analyzes the maximum likelihood estimator for the parameter $\theta_0$ of the above equation and proves its asymptotic normality and moment convergence.
3. **Extension of Theoretical Framework**: The paper extends the field of parameter estimation for stochastic differential equations driven by fractional Brownian motion, particularly in the rough case where the Hurst index $H \in (1/3, 1/2)$, which aligns with situations where traditional semimartingale analysis methods are not applicable.
Through the above work, the paper fills the theoretical gap in parameter estimation under these specific conditions and provides new tools for statistical inference.