131. Astrostatistics in High Energy Astrophysics
7:30pm-9:00pm, 8 Apr 2013, DeAnza Ballroom
This astrostatistics session will honor the
memory of Alanna Connors.
Alanna was a pioneer in the investigation and development of statistical
methods required for the analysis of high energy data. The session
will provide a historical perspective on Bayesian methods, highlight
the topics where the contribution made by Alanna were particularly
important, summarize current analysis challenges and discuss future
perspectives.
Program
- Chair: Herman Marshall (MIT)
-
- Retrospective [20 min]
- 131.01 Alanna Connors and the Origins of Principled Data Analysis
- Jeffrey D. Scargle (NASA/Ames)
- Abstract: Alanna was one of the most important pioneers
in the development of not just sophisticated algorithms for analyzing
astronomical data but more importantly an overall viewpoint
emphasizing the use of statistically sound principles in place
of blind application of cook-book recipes, or black boxes.
I will outline some of the threads of this viewpoint, emphasizing
time series data, with a focus on the importance of these
developments for the Age of Digital Astronomy that we
are entering.
- Presentation slides [.pdf]
-
- Topics [5+2 min]
- 131.02 Promoting widespread use of Bayesian analysis
- Keith Arnaud (NASA/GSFC)
- Abstract: I will discuss incorporating Bayesian analysis in XSPEC
and what is required for them to become the default option for high energy
astrophysicists.
- Presentation notes [.txt]
-
- 131.03 Hypothesis Tests
- Aneta Siemiginowska (CfA)
- Abstract: I will comment on statistical methods for hypothesis
testing, highlighting the potential issues with using posterior predictive
p-values or Bayes Factors.
- Presentation slides [.pdf]
-
- 131.04 Using Bayesian Blocks to Trigger on Very High Energy Flares from Blazars
- James Chiang (Stanford/SLAC)
- Abstract:
Owing to the continuous scanning observation mode of the Fermi
satellite, the Large Area Telescope (LAT) is able to provide all-sky
coverage on 3-hour time scales. However, this observing mode induces
strong exposure modulations for any given location on the sky, making
detection of transients on time scales of minutes to hours from
objects problematic. I will discuss the application of the Bayesian
Blocks algorithm to LAT data in a way that accounts for these exposure
variations and show how the technique might be used to trigger on
bright GeV-TeV-flaring blazars such as Mrk 421.
- Presentation slides [.pdf]
-
- 131.05 A Statistical Approach to Recognizing Source Classes for Unassociated Sources in the Second Fermi-LAT Catalog
- Maria Elena Monzani (Stanford/SLAC)
- Abstract: We have developed a new and innovative technique
to classify Fermi sources based solely on their observed gamma-ray
properties. Our technique, based on Classification Trees, uses the
properties of known objects to build a classification analysis which
provides the probability for an unidentified source to belong to a
given astronomical class (Pulsar, AGN,...). We have applied this
technique to the second Fermi-LAT source catalog (2FGL), and computed
a classification probability for each unidentified source. This provides
a clearer picture of the unidentified source population and extends the
number of interesting candidate objects, thus helping the community in
scheduling multiwavelength observations.
- Presentation slides [.pdf]
- Extended illustration of analysis and methods are given in Ferrara et al. (2012 Fermi & Jansky Proceedings, arXiv/1206.2571)
-
- 131.06 LIRA: Low-counts Image Reconstruction and Analysis
- Nathan M. Stein (Harvard)
- Abstract:
I will discuss LIRA, an R software package for multi-scale nonparametric
image analysis in high energy astrophysics. LIRA uses Markov chain Monte
Carlo to simultaneously fit an image and the necessary smoothing
parameters, allowing for quantification of the standard error of the fitted
image and evaluation of the goodness-of-fit of a proposed model.
- Presentation slides [.pdf]
-
- 131.07 Systematic Errors
- Vinay L. Kashyap (SAO)
- Abstract: I will talk about methods to incorporate
systematic errors into astrostatistical analysis, and discuss
future prospects. In particular, I will point out how the pyBLoCXS
algorithm lends itself to adding feedback, generalize to higher
dimensions, and parameterize the sources of systematics.
- Presentation slides [.pdf]
-
- Prospective [20 min]
- 131.08 Possessing the field: Alanna Connors and the future of principled data analysis
- Thomas J. Loredo (Cornell)
- Abstract:
Alanna emphasized the importance of adopting sound principles to guide
statistical analysis of astronomical data. Her work and the work of her
many collaborators and colleagues demonstrates the value of a principled
approach. I will highlight two threads in her work: hierarchical
Bayesian modeling, and nonparametric Bayesian modeling. I will note how
these powerful ideas have influenced recent astrostatistics research,
and point to how they might profitably merge for addressing the emerging
problems of synoptic time-domain astronomy via Bayesian functional data
analysis.
-
Posters and talks of interest
- 108.10. First Statistical Tests for Clumpy Torii Models: Constraints from RXTE Monitoring of Seyfert AGN
- Alex Markowitz; Mirko Krumpe; Robert Nikutta
- 117.01. X-ray Reflected Spectra from Accretion Disk Models: A Complete Grid of Ionized Reflection Calculations
- Javier Garcia; Thomas Dauser; Christopher S. Reynolds; Timothy R. Kallman; Jeffrey E. McClintock; Ramesh Narayan; Joern Wilms; Wiebke Eikmann
- 117.02. Two-temperature and Model-Independent Differential Emission Measure Distributions: The Emperor's New Clothes?
- Kenneth G. Gayley
- 117.04. Statistical Methods in XSPEC
- Keith A. Arnaud
- 117.05. The HEASARC in 2013 and Beyond: NuSTAR, Astro-H, NICER...
- Stephen A. Drake; Alan P. Smale; Thomas A. McGlynn; Keith A. Arnaud
- 117.06. Application of an Improved Event Reconstruction and Imaging Approach for Compton Telescopes to Crab Measurements by NCT and COMPTEL Using MEGAlib
- Andreas Zoglauer; Steven E. Boggs
- 117.07. Artificial Neural Networks as a Tool to Classify the 2FGL Unassociated Sources
- David Salvetti
- 117.08. Earth Occultation Imaging of the Low Energy Gamma-ray Sky with GBM
- James Rodi; Michael L. Cherry; Gary L. Case; Mark H. Finger; Peter Jenke; Colleen Wilson; Ascension Camero-Arranz; Vandiver Chaplin
- 117.09. Bayesian Methods in Sherpa
- Aneta Siemiginowska; Thomas L. Aldcroft; Vinay Kashyap
- 126.16. Population Synthesis of Radio and Yray Normal, Isolated Pulsars Using Markov Chain Monte Carlo
- Caleb Billman; Peter L. Gonthier; Alice K. Harding
- 127.23. Gamma-Ray Bursts: Pulses and Populations
- Thomas J. Loredo; Jon E. Hakkila; Mary Beth Broadbent; Ira M. Wasserman; Robert L. Wolpert
- 204.03. A Statistical Approach to Identifying Compact Objects in Xray Binaries
- Saeqa D. Vrtilek
Organizers: Aneta Siemiginowska, Vinay Kashyap