Response to Comments on PCA Based Hurst Exponent Estimator for fBm Signals Under Disturbances

Abstract

In this response, we try to give a repair to our previous proof for PCA Based Hurst Exponent Estimator for fBm Signals by using orthogonal projection. Moreover, we answer the question raised recently: If a centered Gaussian process $G_t$ admits two series expansions on different Riesz bases, we may possibly study the asymptotic behavior of one eigenvalue sequence from the knowledge on the asymptotic behaviors of another.