Statistics – Computation
Scientific paper
Dec 2002
adsabs.harvard.edu/cgi-bin/nph-data_query?bibcode=2002spie.4847...35s&link_type=abstract
Astronomical Data Analysis II. Edited by Starck, Jean-Luc; Murtagh, Fionn D. Proceedings of the SPIE, Volume 4847, pp. 35-49
Statistics
Computation
Scientific paper
The future of astronomy will be dominated with large and complex data bases. Megapixel CMB maps, joint analyses of surveys across several wavelengths, as envisioned in the planned National Virtual Observatory (NVO), TByte/day data rate of future surveys (Pan-STARRS) put stringent constraints on future data analysis methods: they have to achieve at least N log N scaling to be viable in the long term. This warrants special attention to computational requirements, which were ignored during the initial development of current analysis tools in favor of statistical optimality. Even an optimal measurement, however, has residual errors due to statistical sample variance. Hence a suboptimal technique with significantly smaller measurement errors than the unavoidable sample variance produces results which are nearly identical to that of a statistically optimal technique. For instance, for analyzing CMB maps, I present a suboptimal alternative, indistinguishable from the standard optimal method with N3 scaling, that can be rendered N log N with a hierarchical representation of the data; a speed up of a trillion times compared to other methods. In this spirit I will present a set of novel algorithms and methods for spatial statistical analyses of future large astronomical data bases, such as galaxy catalogs, megapixel CMB maps, or any point source catalog.
No associations
LandOfFree
Spatial Statistical Analysis of Large Astronomical Datasets 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 Spatial Statistical Analysis of Large Astronomical Datasets, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Spatial Statistical Analysis of Large Astronomical Datasets will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-1310918