Jun 6 – 10, 2016
Europe/Zurich timezone

Techniques for statistical learning including variable transformation, variable selection, and classification are reviewed, with emphasis on low-dimensional datasets typical for physics analysis. The reviewed techniques cover principal component analysis (PCA), kernel PCA, multidimensional scaling, classification by decision trees and their ensembles, classification by kernel methods such as support vector machines, and supervised feature selection by sequential algorithms, ensembles of decision trees and nearest neighbors. Application of these techniques to real-world datasets is illustrated in MATLAB. A MATLAB primer is included in the course, and MATLAB trial licenses will be available to participants.

Attention: The agenda of the course has been changed! Please refer to the timetable to see the updated agenda! The course takes place in the afternoons of Monday, Tuesday, Thursday and Friday. On Wednesday afternoon there is no lecture!

We are happy to announce that the course will be accessible also via vidyo connection. This is possible also for participants that are not registered for the event. Please follow the dedicated link in the menu. 

Due to copyright regulations we cannot make the slides of the course publicly available. If you are not registered for the event, but want to follow the course via vidyo connection you can write to sabine.hemmer(at)pd.infn.it and we will provide you with the slides.  

4/3-006 - TH Conference Room
Go to map

Organized by the AMVA4NewPhysics network, funded by the EC programme H2020 under the Grant Agreement MSCA-ITN-2015-675440

H2020 EU AMVA4NewPhysics CERN