Algorithms for Applied Multivariate Statistical Analysis

点击翻译按钮获取中文摘要

This project concentrates on the development, analysis, and implementation of algorithms for computing the distributions of the eigenvalues (and functions thereof) of the classical random matrix ensembles -- Wishart, Jacobi, and Laguerre. The focus is on achieving high practical efficiency by exploiting the combinatorial and algebraic properties of multivariate orthogonal polynomials as well as utilizing structured matrix computations and dynamic programming techniques.brbrThis research will provide new efficient tools for using multivariate statistical techniques in practice. Such techniques are critical in many areas and applications, including bioinformatics, genomics (population classification), wireless communications (network capacity optimization), and military applications (automatic target classification).