THURSDAY, October 9th, 2008 at 4.10pm,
MSB 1147 (Colloquium Room)
Refreshments: 3.30pm, MSB 4110 (Statistics
Lounge)
Speaker: Randall Eubank (Dept
of
Title: Canonical Correlation and Prediction for Stochastic Processes
Abstract: A general notion of canonical
correlation is developed that expands the classical framework to include
decomposition of cross-covariance structures for stochastic process pairs.
Large sample properties are presented for associated regularized estimators of
the correlations and singular functions and various computational methods are
considered. The relationship between canonical correlation and best linear
unbiased prediction is described and then used to provide a general solution to
a problem of best linear unbiased prediction.
