Event Title

Calculating Multichannel Bayesian Limits Using a Markov Chain Monte Carlo Calculator

Advisor(s)

Dr. Peter Dong, Illinois Mathematics and Science Academy

Dr. Leonard Spiegel, Fermilab

Location

Room A147

Start Date

26-4-2019 11:25 AM

End Date

26-4-2019 11:40 AM

Abstract

Our group's goal is to find evidence of quark-lepton compositeness by analyzing contact interactions that would indicate the presence of preons, theoretical constituents of quarks and leptons. We focus specifically on the Bayesian statistical analysis that determines the lower limit for the energy scale at which such contact interactions would occur. We calculate limits using a Bayesian Markov chain Monte Carlo calculator which utilizes RooStats, a statistical analysis program, to find the 95% confidence interval for a given parameter of interest. We show that we can find simple single-bin limits, then include background processes and systematic uncertainties into the limit calculation before generating multi-channel limits. We are creating a program that computes the Bayesian limit of multiple channels with correlated systematic uncertainties. We will show our promising results and outline the issues that remain to be solved.

Share

COinS
 
Apr 26th, 11:25 AM Apr 26th, 11:40 AM

Calculating Multichannel Bayesian Limits Using a Markov Chain Monte Carlo Calculator

Room A147

Our group's goal is to find evidence of quark-lepton compositeness by analyzing contact interactions that would indicate the presence of preons, theoretical constituents of quarks and leptons. We focus specifically on the Bayesian statistical analysis that determines the lower limit for the energy scale at which such contact interactions would occur. We calculate limits using a Bayesian Markov chain Monte Carlo calculator which utilizes RooStats, a statistical analysis program, to find the 95% confidence interval for a given parameter of interest. We show that we can find simple single-bin limits, then include background processes and systematic uncertainties into the limit calculation before generating multi-channel limits. We are creating a program that computes the Bayesian limit of multiple channels with correlated systematic uncertainties. We will show our promising results and outline the issues that remain to be solved.