Eckhard von Törne: Multivariate methods in statistical data analysis

12:30 -14:30 tutorial (=intro talk + practice session) on multivariate analysis in TMVA 15:00- 16:00 seminar talk on multivariate data analysis

Multivariate analysis is an important component of  particle physics analyses. The toolkit for multivariate analysis, TMVA, provides a large set of advanced multivariate analysis techniques for signal/background classification and regression problems. These techniques can be directly compared in their  performance which allows the user to choose the most appropriate for a particular data analysis. We also discuss the performance of our most important multivariate techniques on example data and a comparison with theoretical performance limits is presented. The application of multivariate analysis in HEP is compared to its usage in other areas.

The seminar is preceded by a tutorial on TMVA, consisting in a basic introduction followed by a hands-on session. Participants in the tutorial should bring a laptop equipped with either

• a working ROOT version (including TMVA as a module)
• ssh + Xwindows support (being able to open XWindows remotely such as xlogo, or ROOT TBrowser) installed.