On the regularity of spectral densities of continuous-time completely linearly regular processes (Q1593596)

Let \(\{X_t\}\) be a real, zero-mean, mean-square continuous, weakly stationary stochastic process with spectral density \(f(\omega)\). For any interval \(I\), let \({\mathcal F}_I\) be the smallest \(\sigma \)-field generated by \(\{X_t, t\in I\}\), and \({\mathcal M}_I\) be the set of \({\mathcal F}_I\)-measurable functions with zero mean and finite second moment. For \(\tau >0\) define the coefficient of complete regularity \[ r(\tau)=\sup \{E(\eta \xi): \eta \in {\mathcal M}_{(-\infty ,0]}, \xi \in {\mathcal M}_{[\tau ,\infty)}, \|\eta \|=\|\xi \|=1\}. \] The process \(\{X_t\}\) is called completely regular if \(r(\tau)\to 0\) as \(\tau \to \infty \). The author proves that when \(r(\tau)=O(\tau ^{-r-\mu})\) for some positive integer \(r\) and \(\mu \in (0,1]\), then \(f(\omega)\) has at least \(r\) uniformly continuous, bounded, and integrable derivatives such that \(f^{(r)}(\omega)\) satisfies Lipschitz continuity condition of order \(\mu \). Conversely, under certain conditions on \(f(\omega)\) upper bounds for \(r(\tau)\) are obtained.