Title: A Semiparametric Approach to Incorporating Systematic Uncertainties
into Bayesian X-ray Spectral Fitting
Authors: Hyunsook Lee, Vinay L. Kashyap, Jeremy J. Drake, Alanna Connors,
Rima Izem, Taeyoung Park, Pete Ratzlaff, Aneta Siemiginowska, David A. van Dyk,
Andreas Zezas
Abstract:
We develop a unique methodology to incorporate systematic uncertainties
into X-ray spectral fitting analysis. These uncertainties have been ignored
in calibrating noisy astronomical data and
as a consequence, error bars of interesting parameters are generally
underestimated. Our strategy combines parametric Bayesian spectral
fitting and nonparametric approximation of the detector characteristics,
the source of systematic uncertainties.  We describe our implementation of
this method here, in the context of recently codified \chandra\ effective area
uncertainties. We estimate the posterior probability densities
of absorbed power-law model parameters that include the effects of systematic
uncertainties.  We apply our method to both simulated as well as actual
\chandra\ ACIS-S data. Because of the modular structure of the Bayesian
spectral fitting technique, incorporating such uncertainties can be
executed simultaneously within the Markov chain Monte Carlo method.
Therefore, our strategy itself does not significantly affect the overall
computing time but offers adequate parameter estimates and error bars.
This research was supported by NASA/AISRP Grant NNG06GF17G.