Brownian Covariance and DISCO Analysis

DATE: October 30, 2009

TIME: 11:00 AM – 12:00 PM

LOCATION: Executive classroom

FEES: none

EVENT CONTACT:

Yajun Mei, ISyE
Contact Yajun Mei
404-894-2300

TITLE: Brownian Covariance and DISCO Analysis

SPEAKER: Dr. Gabor Szekely

ABSTRACT:

The notion of covariance with respect to a stochastic process is introduced. If the process is the identity function, we get the Pearson's covariance, if the process is the Brownian process then we get a new measure of dependence, the Brownian covariance. This is zero if the random variables are independent. The corresponding statistic has an elegantly simple computing formula. DISCO Analysis, that is Distance Components Analysis is a nonparametric extension of ANOVA for testing the multi-sample hypothesis of equal distributions. Many applications will appear in our forthcoming Ann. Appl. Statist. discussion paper.