Parameter identification for the Hermite Ornstein-Uhlenbeck process (Q2194047)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Parameter identification for the Hermite Ornstein-Uhlenbeck process |
scientific article |
Statements
Parameter identification for the Hermite Ornstein-Uhlenbeck process (English)
0 references
25 August 2020
0 references
The authors consider a Hermite Ornstein-Uhlenbeck process \((X_t)_{t\in[0,T]}\) given as the unique solution of the stochastic differential equation (SDE) \[ \mathrm{d} X_t = -a X_t \, \mathrm{d} t + \sigma \, \mathrm{d}Z^{(q,H)}_t,\qquad t\in[0,T], \] where \(T>0\), \(a\in\mathbb{R}\), \(\sigma>0\), the initial condition \(X_0\) is square integrable and \((Z^{(q,H)}_t)_{t\in[0,T]}\) is a Hermite process of order \(q\geq 1\) with Hurst parameter \(H\in(\frac{1}{2},1)\). The authors construct estimators for \(H\) and \(\sigma\) based on discrete time observations of the process \(X\) at times \(\frac{i}{N}\), \(i=0,1,\ldots, N\), where \(N\) is a natural number, and they also analyze their asymptotic behaviour as \(N\to\infty\). A first step to estimate some parameters of a stochastic process given by an SDE driven by a Hermite process is due to [\textit{I. Nourdin} and \textit{T. T. D. Tran}, Stochastic Processes Appl. 129, No. 10, 3774--3791 (2019; Zbl 1479.60078)], where estimators for the drift parameters of a Vasicek-type model driven by a Hermite process based on continuous time observations have been derived. In the present paper, the authors show that their estimator for \(H\) is strongly consistent and they also describe its asympotic behaviour considering two cases when \(\sigma\) is known and not known. Further, they prove that their estimator for \(\sigma\) is weakly consistent provided that \(H\) is known, but the description of its asymptotic behavior remains open. The proposed estimators can be expressed in terms of the quadratic and generalized variations of \(X\), and their limit behaviour is strongly connected to those of the variations in question. In the proofs, the authors distinguish non-stationary and stationary cases, and they use several properties of random variables in Wiener chaos.
0 references
Hermite process
0 references
fractional Brownian motion
0 references
parameter estimation
0 references
multiple Wiener-Itô integrals
0 references
strong consistency
0 references
asymptotic normality
0 references
Ornstein-Uhlenbeck process
0 references
Hurst index estimation
0 references
0 references
0 references
0 references
0 references
0 references
0 references
0 references
0 references