Two approaches to consistent estimation of parameters of mixed fractional Brownian motion with trend (Q2137743)

The authors consider the problem of estimation of parameters \(\theta, \sigma, \kappa\) and the Hurst index \(H\in (0,1)\) from discrete observations of a mixed fractional Brownian motion with trend of the form \(X_t=\theta t +\sigma W_t+ \kappa B_t^H\) where \(W\) is the standard Brownian motion and \(B^H\) is an independent fractional Brownian motion with Hurst index \(H.\) The first method consists of estimation of \(\sigma,\kappa\) and \(H\) using the quadratic variation and the estimation of the parameter \(\theta\) by discretization of the continuous-time maximum likelihood estimator. This method is applicable when \(H <\frac{3}{4}.\) The second method uses the ergodic theorem and it is applicable for all \(H\in (0,1).\) The methods are compared using Monte Carlo simulations.