Seminar in applied mathematics and statistics

SPEAKER: Line Katrine Harder Clemmensen (DTU)

TITLE: Slimming big data

ABSTRACT:
Framed within big data analytics, this talk will address methods for predictive analytics when we face problems with many variables and few observations. In particular, we will look at sparse discriminant analysis and random forest models, and consider interpretations in such high-dimensional spaces. For problems with many variables it often holds that we have both redundant and irrelevant variables; shrinkage as well as variable selection are useful to deal with these types of data, and linking predictive ability to domain knowledge is crucial. The talk will discuss these aspects of the models in question.

Tea and chocolate will be served in room 04.3.15 after the seminar.

-----

CURRENT SCHEDULE FOR SPRING 2016:

January 20, 15.15, Aud. 6: Alessia Pini (MOX, Politecnico di Milano)
February 24, 15.15, Aud. 10: Matthias Fahrenwaldt (Hannover)
March 9, 15.15, Aud. 10: Mogens Fosgerau (DTU)
March 11, 14.15, Aud. 8: Hansjoerg Albrecher (Lausanne)
March 16, 15.15, Aud. 10: Peter McCullagh (Oxford/Chicago)
March 17, 15.15, Aud. 10: Peter Glynn (Stanford)
May 13, 14.15, Aud. 3: Emilio Procu (Univ. Federico Santa Maria, Chile)
May 18, 15.15, Aud. 8: Holger Drees (Hamburg)
May 20, 14.15, Aud. 10: Kim Christensen (Aarhus)
June 8, 15.15, Aud. 8: Ana-Maria Staicu
June 15, 15.15, Aud. 8: Line K. H. Clemmensen