F. Tegenfeldt

TMVA : The Toolkit for Multivariate Data Analysis eith ROOT

Multivariate classification methods based on machine learning techniques play a fundamental role in today’s high-energy physics analyses dealing with ever smaller signal in ever larger data sets. TMVA is a toolkit integrated in the ROOT framework which implements a large variety of multivariate classification algorithms ranging from simple rectangular cut optimisation and likelihood estimators, over linear and non-linear discriminants to more recent developments like boosted decision trees, rule fitting and support vector machines. All classifiers can be trained, tested and evaluated simultaneously. They all see the same training are then afterwards also tested on the same independent test data allowing meaningful comparisons between the methods for a particular use case. Here, an overview about the package and the classifiers currently implemented is presented.

Operation, optimisation, and performance of the DELPHI RICH detectors

The Ring Imaging Cherenkov detectors of DELPHI represent a large-scale particle identification system which covers almost the full angular acceptance of DELPHI. The combination of liquid and gas radiators (C4F10, C5F12, and C6F14) provides particle identification over the whole secondary particle momentum spectrum at LEP I and LEP II. Continuing optimisation on the hardware as well as on the online and offline software level have resulted in a stable operation of the complete detector system for more than five years at full physics performance.

On Bayesian Treatment of Systematic Uncertainties in Confidence Interval Calculation

In high energy physics, a widely used method to treat systematic uncertainties in confidence interval calculations is based on combining a frequentist construction of confidence belts with a Bayesian treatment of systematic uncertainties. In this note we present a study of the coverage of this method for the standard Likelihood Ratio (aka Feldman & Cousins) construction for a Poisson process with known background and Gaussian or log-Normal distributed uncertainties in the background or signal efficiency. For uncertainties in the signal efficiency of upto 4 0 % we find over-coverage on the level of 2 to 4 % depending on the size of uncertainties and the region in signal space. Uncertainties in the background generally have smaller effect on the coverage. A considerable smoothing of the coverage curves is observed. A software package is presented which allows fast calculation of the confidence intervals for a variety of assumptions on shape and size of systematic uncertainties for different nuisance parameters. The calculation speed allows experimenters to test the coverage for their specific conditions.

Single intermediate vector boson production in e(+)e(-) collisions at root s=183-209 GeV

Abstract.The production of single charged and neutral intermediate vector bosons in e + e- collisions has been studied in the data collected by the DELPHI experiment at LEP at centre-of-mass energies between 183 and 209 GeV, corresponding to an integrated luminosity of about 640 pb-1. The measured cross-sections for the reactions, determined in limited kinematic regions, are in agreement with the Standard Model predictions.