Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userA model independent safeguard for unbinned Likelihood
84
0
0.0
(
0
)
Ask ChatGPT about the research
No Arabic abstract
We present a universal method to include residual un-modeled background shape uncertainties in likelihood based statistical tests for high energy physics and astroparticle physics. This approach provides a simple and natural protection against mismodeling, thus lowering the chances of a false discovery or of an over constrained confidence interval, and allows a natural transition to unbinned space. Unbinned likelihood allows optimal usage of information for the data and the models, and enhances the sensitivity. We show that the asymptotic behavior of the test statistic can be regained in cases where the model fails to describe the true background behavior, and present 1D and 2D case studies for model-driven and data-driven background models. The resulting penalty on sensitivities follows the actual discrepancy between the data and the models, and is asymptotically reduced to zero with increasing knowledge.
rate research
Read More
PyUnfold is a Python package for incorporating imperfections of the measurement process into a data analysis pipeline. In an ideal world, we would have access to the perfect detector: an apparatus that makes no error in measuring a desired quantity. However, in real life, detectors have finite resolutions, characteristic biases that cannot be eliminated, less than full detection efficiencies, and statistical and systematic uncertainties. By building a matrix that encodes a detectors smearing of the desired true quantity into the measured observable(s), a deconvolution can be performed that provides an estimate of the true variable. This deconvolution process is known as unfolding. The unfolding method implemented in PyUnfold accomplishes this deconvolution via an iterative procedure, providing results based on physical expectations of the desired quantity. Furthermore, tedious book-keeping for both statistical and systematic errors produces precise final uncertainty estimates.
LISA is the upcoming space-based Gravitational Wave telescope. LISA Pathfinder, to be launched in the coming years, will prove and verify the detection principle of the fundamental Doppler link of LISA on a flight hardware identical in design to that of LISA. LISA Pathfinder will collect a picture of all noise disturbances possibly affecting LISA, achieving the unprecedented pureness of geodesic motion necessary for the detection of gravitational waves. The first steps of both missions will crucially depend on a very precise calibration of the key system parameters. Moreover, robust parameters estimation is of fundamental importance in the correct assessment of the residual force noise, an essential part of the data processing for LISA. In this paper we present a maximum likelihood parameter estimation technique in time domain being devised for this calibration and show its proficiency on simulated data and validation through Monte Carlo realizations of independent noise runs. We discuss its robustness to non-standard scenarios possibly arising during the real-life mission, as well as its independence to the initial guess and non-gaussianities. Furthermore, we apply the same technique to data produced in mission-like fashion during operational exercises with a realistic simulator provided by ESA.
We propose a novel method for computing $p$-values based on nested sampling (NS) applied to the sampling space rather than the parameter space of the problem, in contrast to its usage in Bayesian computation. The computational cost of NS scales as $log^2{1/p}$, which compares favorably to the $1/p$ scaling for Monte Carlo (MC) simulations. For significances greater than about $4sigma$ in both a toy problem and a simplified resonance search, we show that NS requires orders of magnitude fewer simulations than ordinary MC estimates. This is particularly relevant for high-energy physics, which adopts a $5sigma$ gold standard for discovery. We conclude with remarks on new connections between Bayesian and frequentist computation and possibilities for tuning NS implementations for still better performance in this setting.
The current and upcoming generation of Very Large Volume Neutrino Telescopes---collecting unprecedented quantities of neutrino events---can be used to explore subtle effects in oscillation physics, such as (but not restricted to) the neutrino mass ordering. The sensitivity of an experiment to these effects can be estimated from Monte Carlo simulations. With the high number of events that will be collected, there is a trade-off between the computational expense of running such simulations and the inherent statistical uncertainty in the determined values. In such a scenario, it becomes impractical to produce and use adequately-sized sets of simulated events with traditional methods, such as Monte Carlo weighting. In this work we present a staged approach to the generation of binned event distributions in order to overcome these challenges. By combining multiple integration and smoothing techniques which address limited statistics from simulation it arrives at reliable analysis results using modest computational resources.
We consider whether the asymptotic distributions for the log-likelihood ratio test statistic are expected to be Gaussian or chi-squared. Two straightforward examples provide insight on the difference.
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
