Repository logo
 

Statistical methods for the detection and analysis of radioactive sources

Date

2014

Authors

Klumpp, John, author
Brandl, Alexander, advisor
Johnson, Thomas, committee member
Steinhauser, Georg, committee member
LaRue, Susan, committee member
Givens, Geof, committee member

Journal Title

Journal ISSN

Volume Title

Abstract

We consider four topics from areas of radioactive statistical analysis in the present study: Bayesian methods for the analysis of count rate data, analysis of energy data, a model for non-constant background count rate distributions, and a zero-inflated model of the sample count rate. The study begins with a review of Bayesian statistics and techniques for analyzing count rate data. Next, we consider a novel system for incorporating energy information into count rate measurements which searches for elevated count rates in multiple energy regions simultaneously. The system analyzes time-interval data in real time to sequentially update a probability distribution for the sample count rate. We then consider a "moving target" model of background radiation in which the instantaneous background count rate is a function of time, rather than being fixed. Unlike the sequential update system, this model assumes a large body of pre-existing data which can be analyzed retrospectively. Finally, we propose a novel Bayesian technique which allows for simultaneous source detection and count rate analysis. This technique is fully compatible with, but independent of, the sequential update system and moving target model.

Description

Rights Access

Subject

Citation

Associated Publications