Statistics – Computation
Scientific paper
Dec 2001
adsabs.harvard.edu/cgi-bin/nph-data_query?bibcode=2001aas...19911201v&link_type=abstract
American Astronomical Society, 199th AAS Meeting, #112.01; Bulletin of the American Astronomical Society, Vol. 33, p.1477
Statistics
Computation
Scientific paper
Please join us in our wider effort to engage the strengths of modern computational statistics methods in solving challenging stellar and solar data analysis problems. As just one example (of a great breadth of possibilities) consider analyzing a spectrum with a very large number of lines. Some of these may be faint, merged, indistinguishable from each other and the underlying smooth continuum. The ensemble of line intensities follows a predictable distribution. The shape of this distribution depends on the properties of the source, e.g., its temperature, abundances, and emission measure. Hence, a better understanding of the distribution of line fluxes in a particular source may tighten our inference for other model parameters such as temperature---even when very few lines are actually easy to distinguish. To take advantage of this structure, we directly model the distribution of the line fluxes rather than fitting each line flux directly or ``investing'' the emissivities to get a DEM. Statistically, this strategy reduces the number of free parameters, which we expect will lead to improved statistical properties. We believe this method holds much promise for improved analysis, especially for low count sources. For example, we expect this method to correctly account for the ``pseudo-continuum'' that results from the large number of faint, unresolvable lines in X-ray grating spectra. Moreover, our statistical methods should apply directly to other settings involving a multitude of lines such as timing data. We hope that these methods will increase our statistical power to set the continuum level in the presence of a multitude of lines and to distinguish weak lines from fluctuations in the continuum. Funding for this project partially provided by NSF grant and DMS-01-04129 and by NASA Contract NAS8-39073 (CXC).
Connors Alanna
Kang Hyung Suk
Kashyap Vinay L.
Siemiginowska Aneta
van Dyk David A.
No associations
LandOfFree
Statistical Analysis of Spectra with Many Lines does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with Statistical Analysis of Spectra with Many Lines, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Statistical Analysis of Spectra with Many Lines will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-1236075