Synthesis of Radiation Counts for Networks of Detectors

We develop a data-driven, regression-based scaling method to synthesize radiation counts for a given source strength in a spectral bin by bootstrapping measurements from Domestic Nuclear Detection Office's Intelligence Radiation Sensors Systems (IRSS) tests. In addition to counts, we also generate network messages using message-skeletons extracted from IRSS tests. We characterize the quality of generated counts using statistical mea- sures and distance metrics, and provide an analytical justification for our method based on the Vapnik-Chervonenkis theory.