On the asymptotic behaviour of the ε,δ-entropy of the system of Gaussian measures inRn
Let \(\Lambda\) be a set of probability measures on a measurable space (X,\({\mathfrak A})\), \(X_ 0\subset X\) and \(\forall \mu \in \Lambda:\) \(\mu X_ 0=1\); \(H_{\epsilon}(A,X)\) is the \(\epsilon\)-entropy of a set A in \((X_ 0,\rho)\), where \(\rho\) is a metric on \(X_ 0\). The author defines the \(\epsilon\),\(\delta\)-entropy \[ H_{\epsilon,\delta}(\Lambda,X)=\inf \{H_{\epsilon}(X_ 0\setminus A,X_ 0):\quad A\in {\mathfrak A},\quad \mu A\leq \delta,\quad \forall \mu \in \Lambda \}. \] In this paper it is proved that if \(\Lambda_ n(r)\) is a system of Gaussian measures in R, then \(\forall \epsilon >0\), \(1>\delta >0\), \(r>0:\) \[ \lim_{n\to \infty}n^{-1} H_{\epsilon,\delta}(\Lambda_ n(r),R^ n)=(\log r\epsilon^{-1})_+. \]
- scientific article; zbMATH DE number 1530003
- Approximation, metric entropy and small ball estimates for Gaussian measures
- On the<tex>epsilon</tex>-entropy of certain Gaussian processes
- Metric entropy and the small ball problem for Gaussian measures
- Sharp asymptotics of the Kolmogorov entropy for Gaussian measures
This page was built for publication: On the asymptotic behaviour of the ε,δ-entropy of the system of Gaussian measures inRn
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3683240)