Korbinian Strimmer
Konstantinos Fokianos
Summer term 2005


Strategies for Analyzing High-Dimensional Data.

In many current problems in the biosciences (genomics, proteomics
etc.) but also in astronomy and econometrics one encounters data sets
of extremely high dimension (p) that cannot be analyzed with
traditional statistical tools due to small sample size (n). Using
examples from recent literature we will discuss suggestions and
strategies of how to best deal with these "small n, large p" data.
Topics will include empirical Bayes inference, regularization,
large-scale multiple testing, small-sample classification, graphical
models, time series analysis etc.

The seminar takes place on Monday 10-12, Seminarraum, Ludwigstr 33,
and is coorganized by Konstantinos Fokianos (Cyprus University).

First meeting: 18 April 2005