Mathematics Colloquia and Seminars

Return to Colloquia & Seminar listing

Largest Eigenvalues and Eigenvectors in Multivariate Statistical Analysis


Speaker: Iain M. Johnstone, Department of Statistics, Stanford University
Location: 1147 MSB
Start time: Fri, May 28 2010, 4:10PM

The eigenvalues of Wishart matrices play a central role in classical multivariate statistical analysis. A new impetus to approximate distribution results has come from methods that imagine the number of variables as large. We focus on the largest eigenvalue in particular, and review null distribution approximations to Gaussian, single and double Wishart problems in terms of the Tracy-Widom laws. If time permits, we will also briefly mention estimation of the eigenvectors associated to the top eigenvalues.