scientific article; zbMATH DE number 1098750
For a sample \( X_1, \ldots, X_n \) of iid random vectors one can define the range in different ways. The author takes the univariate range \( R_n(\theta) \) of the projected random points \( X_k(\theta) \) on the line with a certain angle \( \theta \) . In the case of iid normal random vectors in \( R^2 \) which are standardized with correlation coefficient \( \rho, |\rho |< 1 \), the limit behaviour of the stochastic process \[ \xi_n(t):=b_n \{ R_n(\theta+ \Delta_n t)/\sqrt {1+ \rho \sin 2(\theta+ \Delta_n t) }- 2b_n \} \] describes the behaviour of all directions. The main result states that the convergence of \(\frac{\Delta_n^2 (1- \rho^2 )}{2(1+ \rho \sin 2\theta )^2} \log n\) is necessary and sufficient for the convergence \(\xi_n(t) \Rightarrow \xi (t)\) in \( C(-T,T), n \rightarrow \infty, T>0\). For the entire collection see [Zbl 0878.00066].
- Bounds on the Bivariate Normal Distribution Function
- Some characterizations of the bivariate normal distribution
- Normal Vector of a Random Hyperplane
- A Note on Bivariate Normal Distribution
- On equivalence of two variances of a bivariate normal vector
- Correlations between random projections and the bivariate normal
This page was built for publication:
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4368922)
