Applications of Likelihood Analysis in Gamma-Ray Astrophysics

Abstract

The field of gamma ray astronomy relies heavily on the statistical analysis of data. Because of the paucity of data, and the often large errors associated with detecting gamma rays, analysis and interpretation of the data require sophisticated statistical techniques. Techniques for extending the currently used maximum likelihood technique to more complicated data sets are presented. Similarly presented are methods of calculating the distribution and behavior of the maximum likelihood statistic used to measure source significance. A new method for calculating source variability is also proposed, and used to examine the sources found by the Energetic Gamma Ray Experiment Telescope (EGRET) on board the Compton Gamma Ray Observatory (CGRO). The results show that the Active Galactic Nuclei (AGN), pulsars, and unidentified sources have markedly different variability, and that the unidentified sources fall into at least two classes, differing in variability and spatial distribution. A class composed of possible Supernova Remnant associations (SNR) is distinctly more variable than the pulsars, but has a variability consistent with that of the other low latitude unidentified sources. Finally, the results of the Fall 1997 GLAST prototype beam test are presented. The results of the beam test are compared with simulated results, and found to be in remarkably good agreement.