High Dimensional Data and Multivariate Analysis (NDNS+)

Credits 8 credit points
Instructors Jacqueline Meulman (Universiteit Leiden), Aad van der Vaart (Vrije Universiteit), Mark van de Wiel (Vrije Universiteit)
E-mail jmeulman at math.leidenuniv.nl, aad at few.vu.nl, mark.vdwiel at vumc.nl
Description This course gives an overview of techniques for analysing high-dimensional data, e.g. arising from microarray experiments, mass spectronomy, or high-throughput genotyping, including some statistical theory about the quality of such procedures. Among the topics may be regression, classification, multiple testing, and clustering, from the point of view of statistics and statistical learning. The methods discussed will include support vector machines, regression trees, boosting, model selection methods.
Organization Lectures, reading, and possibly presentations by the participants.
Examination Project or oral exam. Project assignment with supporting material
Literature
Prerequisites No specific requirements.

COURSE SPRING 2008
Meetings Wednesdays 14-17, Room 0.17, Mathematics-UvA, Plantage Muidergracht 24.
From 6 February until 21 May, with the exception of March 5, March 26 and April 30.
Schedule 6feb Aad van der Vaart chap 1+2 of HTF
13feb Aad van der Vaart chap 3 of HTF
20feb Aad van der Vaart chap 4+12 of HTF
27feb Aad van der Vaart support vector machines
12 mar Mark van de Wiel micro-array analysis and
19mar Mark van de Wiel multiple testing
2ap Jacqueline Meulmanchap9 HTF
9ap Jacqueline Meulman chap10 HTF
16ap Jacqueline Meulmanchap14 HTF
23ap Jacqueline Meulmangeneralized additive models
7may Jacqueline Meulman clustering
14may Aad van der Vaartchap 5+6 HTF (splines, kernel and polynomial smoothing)
21may Aad van der Vaart theory of minimum contrast, structured risk minimization