Bill Tompkins

I was a grad student working on EGRET data analysis, working on Maximum Likelihood methods of finding point sources, and issues of source variability and source distributions.

My thesis (available in PostScript or Gzipped PostScript ) is titled "Applications of Likelihood Analysis in Gamma-Ray Astrophysics", and deals with an analysis of the variability of the EGRET sources, the Fall '97 GLAST beamtest, and some more general issues regarding Likelihood analysis.

Published Papers:

• EGRET Gamma-Ray Observations of the Crab P2/P1 Ratio (ApJ, Sept 20 1997)
  PS (154K) | compressed PS (75K)

Writeups:

The following are discussions of original work which I have not tried to turn into real papers. As such, they're a bit rough, for which I apologize. I'd be happy to discuss the work, and would appreciate any feedback about the writeups (errors, typos, difficulties, extensions, etc...)

• Why TS is Not (Necessarily) Distributed Like Chi^2 with 3 D.O.F ( DVI , or PS )
  A discussion of why Wilks' theorem does not apply to the distribution of the EGRET Test Statistic when there is no source.

• Notes on the TS statistic used in Maximum Likelihood Analysis of Gamma Ray Telescope Data ( PS )
  A derivation of the distribution of TS when no source is present.

• Computation of Expected TS and Thresholds for EGRET ( PS )
  Presents threshold calculations with a fair amount of detail, but does not discuss skew of the source TS distribution or applications.

• Results of Computation of Expected TS and Thresholds for EGRET ( PS )
  Summarizes earlier work, deals with skew of the source TS distribution and provides empirical fits to the results.