Eigenvector distribution of Wigner matrices (Q1950383): Difference between revisions

Eigenvector distributions of Hermitian and random symmetric \(N \times N\) matrices are studied. Based on the method by \textit{L. Erdős, H.-T. Yau} and \textit{J. Yin} [Adv. Math. 229, No. 3, 1435--1515 (2012; Zbl 1238.15017)], the edge universality is extended to eigenvectors associated with eigenvalues close to the spectral edge. Under the assumption of matching of the first two moments of matrix entries, it is proved that the edge eigenvectors of Hermitian and symmetric Wigner matrices show the same joint distribution of the eigenvectors as those of corresponding Gaussian ensembles. Applying the level repulsion condition, a more general result is proved saying that the joint eigenvector-eigenvalue distribution of two generalized Wigner matrix ensembles coincides near the spectral edge if the first two moments of the Wigner ensembles match. Eigenvectors associated with bulk eigenvalues have also the same joint distribution of the eigenvectors as those of Gaussian ensembles under the condition that the first four moments of the distributions match with the moments of the corresponding Gaussian ensembles.