Secondary Logo

Journal Logo

Institutional members access full text with Ovid®

Simultaneous Source Detection and Analysis Using a Zero-inflated Count Rate Model

Klumpp, John; Brandl, Alexander*

doi: 10.1097/HP.0000000000000291
Papers
Buy

This paper proposes a novel Bayesian technique that allows for simultaneous source detection and count rate analysis. The technique involves using priors, which include a finite probability that the source count rate is exactly zero. Such priors are called “zero-inflated.” Solving the posterior distribution of a zero-inflated count rate model provides the probability that the sample contains a source and a probability distribution for the source count rate if the source exists, without the need to perform redundant computations. Sampling from zero-inflated distributions is straightforward and can be accomplished with easily accessible open source software. In addition, zero-inflated priors lead to finite posterior probabilities of “no source,” which is an easy-to-understand and satisfying result.

Colorado State University, 399 MRB. 1681 Campus Delivery, Fort Collins, CO 80523-1681.

The authors declare no conflicts of interest.

For correspondence contact: John Klumpp, Los Alamos National Laboratory, Los Alamos, NM 87545, or email at jaklumpp@gmail.com.

(Manuscript accepted 16 February 2015)

© 2015 by the Health Physics Society