|
|
|
Scalable Algorithms for Analysis of Megapixel CMB Maps and Large Databases
|
|
|
| We propose to develop a set of algorithms for spatial statistical analyses of large astronomical data bases, such a
CMB maps, galaxy catalogs, or any point source catalog. We will build on the success of our previous AISR, and
tackle computational challenges such as
i) fast and accurate estimation of temperature,
polarization and lensing correlation functions and
power spectra, ii) fast estimation of higher order
correlation functions from large CMB maps,
angular, redshift and weak lensing surveys.
iii) fast estimation of covariance matrices using novel Monte Carlo techniques.
In our unique interdisciplinary approach for algorithmic
development, we reformulate the statistical problems arising
in space astronomy with special attention to computational needs. Our philosophy naturally blends the principles
of astronomy, statistics, computer and computational science.
Special attention will be payed
to the user interface and quality control to ensure wide
practical usage. The resulting software package will be useful for accurate analysis of MAP and Planck, or any subsequent
megapixel surveys, galaxy surveys, background light correlations, correlations in maps produced in other wavelengths, such as infrared background, as well as cross correlations of all the above. |
|
|
