• Corpus ID: 117066810

On hypothesis testing, trials factor, hypertests and the BumpHunter

@article{Choudalakis2011OnHT,
  title={On hypothesis testing, trials factor, hypertests and the BumpHunter},
  author={Georgios Choudalakis},
  journal={arXiv: Data Analysis, Statistics and Probability},
  year={2011}
}
  • G. Choudalakis
  • Published 2 January 2011
  • Physics
  • arXiv: Data Analysis, Statistics and Probability
A detailed presentation of hypothesis testing is given. The "look elsewhere" effect is illustrated, and a treatment of the trials factor is proposed with the introduction of hypothesis hypertests. An example of such a hypertest is presented, named BumpHunter, which is used in the recent ATLAS dijet resonance search, and in an earlier version in the CDF Global Search, to look for exotic phenomena in high energy physics. As a demonstration, the BumpHunter is used to address Problem 1 of the Banff… 
Learning new physics from an imperfect machine
We show how to deal with uncertainties on the Standard Model predictions in an agnostic new physics search strategy that exploits artificial neural networks. Our approach builds directly on the
CWoLa Hunting: Extending the Bump Hunt with Machine Learning
TLDR
This work presents a new method for using all of the available information (with machine learning) without using any prior knowledge about potential signals for a model independent search for new particles at the Large Hadron Collider.
submitter : Study of Alternative Statistical Methods for a Model Unspecific Search in CMS
TLDR
This thesis extends the MUSiC analysis by two components, and proposes a different algorithm, called “Bump Hunter”, a purely data driven analysis, searching for deviations in the smooth distribution of certain observables, which doesn’t rely on Standard Model Monte Carlo predictions.
Search for extra dimensions in the diphoton final state with ATLAS
TLDR
The large difference between the Planck scale and the electroweak scale, known as the hierarchy problem, has been addressed in some models through the existence of extra spatial dimensions, but an analysis of the diphoton final state in data recorded with the ATLAS detector at the CERN Large Hadron Collider finds no evidence ofextra spatial dimensions.
Learning new physics from a machine
TLDR
This work proposes using neural networks to detect data departures from a given reference model, with no prior bias on the nature of the new physics responsible for the discrepancy, and constructs an algorithm that implements this approach, as a straightforward application of the likelihood-ratio hypothesis test.
Guiding new physics searches with unsupervised learning
TLDR
A statistical test upon a test statistic which measures deviations between two samples, using a Nearest Neighbors approach to estimate the local ratio of the density of points, which is model-independent and non-parametric.
Extending the Bump Hunt with Machine Learning
TLDR
A new extension of the bump hunt that naturally benefits from modern machine learning algorithms while remaining model-agnostic is presented, based on the Classification Without Labels (CWoLa) method where the invariant mass is used to create two potentially mixed samples, one with little or no signal and one with a potential resonance.
An Imperfect machine to search for New Physics: systematic uncertainties in a machine-learning based signal extraction
We show how to deal with uncertainties on the Standard Model predictions in an agnostic new physics search strategy exploiting artificial neural networks. Our approach builds directly on the Maximum
Search for high-mass diphoton resonances in pp collisions at √s = 8 TeV with the ATLAS detector
This article describes a search for high-mass resonances decaying to a pair of photons using a sample of 20.3  fb−¹ of pp collisions at √s = 8 TeV recorded with the ATLAS detector at the Large Hadron
RPV stops bump off the background
We study the 8 TeV LHC reach on pair produced heavy flavored di-jet resonances. Motivated by theories of R-parity violation in supersymmetry we concentrate on a final state with two b-jets and two
...
...

References

SHOWING 1-10 OF 12 REFERENCES
Trial factors for the look elsewhere effect in high energy physics
When searching for a new resonance somewhere in a possible mass range, the significance of observing a local excess of events must take into account the probability of observing such an excess
Sleuth: A quasi-model-independent search strategy for new physics
How can we search for new physics when we only vaguely know what it should look like? How can we perform an unbiased yet data-driven search? If we see apparently anomalous events in our data, how can
Global search for new physics with 2.0 fb(-1) at CDF
Data collected in run II of the Fermilab Tevatron are searched for indications of new electroweak-scale physics. Rather than focusing on particular new physics scenarios, CDF data are analyzed for
Sleuth at CDF, a quasi-model-independent search for new electroweak scale physics
These proceedings describe Sleuth, a quasi-model-independent search strategy targeting new electroweak scale physics, and its application to 927 pb^-1 of CDF II data. Exclusive final states are
Model-independent and quasi-model-independent search for new physics at CDF
Data collected in run II of the Fermilab Tevatron are searched for indications of new electroweak scale physics. Rather than focusing on particular new physics scenarios, CDF data are analyzed for
A General search for new phenomena in ep scattering at HERA
A model-independent search for deviations from the Standard Model prediction has been performed in e+p and e–p collisions at HERA using H1 data. All experimentally measurable event topologies
Search for new particles in two-jet final states in 7 TeV proton-proton collisions with the ATLAS detector at the LHC.
TLDR
A search for new heavy particles manifested as resonances in two-jet final states in 7 TeV proton-proton collisions by the LHC is presented, extending the reach of previous experiments.
The RooStats Project
TLDR
This research presents a meta-analysis of proton-proton collisions at CERN using a platform that combines X-ray diffraction, radiolysis, and high-resolution 3D image analysis to characterize the geometry of the nucleus.
ROOT — An object oriented data analysis framework
Search for New Particles in Two-Jet Final S tates in 7 TeV Proton-Proton Collisions with the ATLAS Detector at the LHC
  • Physical Review Letters 105no. 16, (Oct., 2010) 161801, arXiv:1008.2461 [hep-ex].
  • 2010
...
...