PhyStat2007

Contribution List

4. A Bayesian approach to the Constrained MSSM

Prof. Leszek Roszkowski (CERN and University of Sheffield)

We employ a Markov Chain Monte Carlo scanning technique and a Bayesian analysis to perform efficient parameter inference of the CMSSM (Constrained Minimal Supersymmetric Standard Model). The approach allows us to vary simultaneously all the CMSSM parameters and relevant Standard Model (nuisance) parameters, and to properly treat experimental constraints from collider physics and...

12. A pitfall in estimating contributions of systematic errors

James Linnemann (Michigan State University)

A common practice in evaluating the contribution of various systematic uncertainties to a result is to evaluate the contribution of each uncertainty separately (typically changing the parameter by what are felt to be + and - one sigma of systematic uncertainty), then to add the induced changes in the result in quadrature. Statisticians would refer to this as "One Factor at a Time"...

18. Combining Channels with Fits

William Quayle (Wisconsin)

We study a combined fit of several channels in a hypothetical search for new physics. In particular, we use toy Monte Carlo to investigate the potential advantages and disadvantages of imposing consistency requirements, such as demanding that the standalone fits in the individual channels yield masses that are compatible with each other before a combined fit is performed.

3. Confidence distributions in statistical inference

Dr Sergey Bityukov (INSTITUTE FOR HIGH ENERGY PHYSICS, PROTVINO)

This report reviews new methodology for statistical inferences. Point estimators, confidence intervals and $p-$values have been fundamental tools for frequentist statisticians. Confidence distributions, which can be viewed as ``distribution estimators'', are often convenient devices for constructing all the above statistical procedures plus more.

14. Dilution of a statistical significance of a signal in the Higgs boson searches in the H->ZZ channel at LHC

Dr Alexey Drozdetskiy (University of Florida, Gainesville, FL, USA)

Should an event excess compatible with the H->ZZ->4l decay channel be observed at LHC, the statistical significance of the access must be properly scaled down to account for the systematic errors and the fact that the search is performed in a wide-open range of possible Higgs boson masses. In this talk, we present results of studies addressing both of the two contributions and show that...

7. Evaluation of two methods for incorporating a systematic uncertainty into a test of the background-only hypothesis for a Poisson process

Jordan Tucker (UCLA)

Hypothesis tests for the presence of new sources of Poisson counts amidst background processes are frequently performed in high energy physics, gamma ray astronomy, and other branches of science. While there are conceptual issues already when the mean rate of background is precisely known, the issues are even more difficult when the mean background rate has non-negligible uncertainty, as...

11. Event Weighting with near-optimal variance for background-subtracted data

James Linnemann (Michigan State University)

It is possible to find event weighting schemes which produce parameter estimates with variance nearly the same as a ML estimate. But there are situations in which a full ML estimate is inconvenient, usually for computational reasons (iteration over large data sets for example). If an variable x associated with the events is a candidate discriminating variable (that is, its distribution for...

5. Probability Matching Priors in LHC Physics: A Pragmatic Approach

Mr Paul Baines (Harvard University)

Probability matching priors (PMP's) provide a bridge between Bayesian and Frequentist inference by yielding Bayesian posterior intervals with Frequentist validity. PMP's also allow the Frequentist access to the powerful computational tools of Bayesian methodology. Unfortunately, such priors are, in general, extremely challenging to implement as they are defined as the solution to a potentially...

16. SFitter: Determining supersymmetric parameters

Dr Remi Lafaye (LAPP Annecy)

If supersymmetry (or similar complex models) is found at the LHC, the goal for all colliders over the coming decades will be to extract its fundamental parameters from the measurements. Dedicated state-of-the-art tools will be necessary to link a wealth of measurements to an e.g. 20-dimensional MSSM parameter space. Starting from a general log-likelihood function of this...

2. Statistical bias of fit parameters

Sergei Redin (Budker Institute)

In the presentation we consider systematic biases of fit parameters arising from a chi^2 minimization procedure, which we call "statistical biases of fit parameters". We discuss several possible techniques, which may reduce those statistical biases. That may be extremely useful, in paticular for precision experiments with high statistics, if for technical reasons, for systematic...

10. Statistical Software in ROOT

Dr Lorenzo Moneta (CERN)

Advanced mathematical and statistical computational methods are required by the LHC experiments for analyzing their data. Some of these methods are provided by the ROOT project, a C++ Object Oriented framework for large scale data handling applications. Various statistical classes and methods currently exist in ROOT spread in various libraries. Examples include methods for regression...

17. Subtracting histograms using profile likelihood

Mr Andre Nepomuceno (Federal University of Rio de Janeiro)

It is known that many interesting signals expected at LHC are of unknown shape and strongly contaminated by background events. These signals will be difficult to detect during the first years of LHC operation due to the initial low luminosity. In this work, one proposes a method on how to obtain signal information of unknow shape from data even when there are very low signal and large...

1. TMVA - Toolkit for Multivariate Data Analysis with ROOT

Dr Fredrik Tegenfeldt (Iowa State University)

In high-energy physics, with the search for ever smaller signals in ever larger data sets, it has become essential to extract a maximum of the available information from the data. Multivariate classification methods based on machine learning techniques have become a fundamental ingredient to most analyses. Also the multivariate classifiers themselves have significantly evolved in...