J. D. McEwen, S. M. Feeney, M. C. Johnson, H. V. Peiris
A number of well-motivated extensions of the LCDM concordance cosmological
model postulate the existence of a population of sources embedded in the cosmic
microwave background (CMB). One such example is the signature of cosmic bubble
collisions which arise in models of eternal inflation. The most unambiguous way
to test these scenarios is to evaluate the full posterior probability
distribution of the global parameters defining the theory; however, a direct
evaluation is computationally impractical on large datasets, such as those
obtained by the Wilkinson Microwave Anisotropy Probe (WMAP) and Planck. A
method to approximate the full posterior has been developed recently, which
requires as an input a set of candidate sources which are most likely to give
the largest contribution to the likelihood. In this article, we present an
improved algorithm for detecting candidate sources using optimal filters, and
apply it to detect candidate bubble collision signatures in WMAP 7-year
observations. We show both theoretically and through simulations that this
algorithm provides an enhancement in sensitivity over previous methods by a
factor of approximately two. Moreover, no other filter-based approach can
provide a superior enhancement of these signatures. Applying our algorithm to
WMAP 7-year observations, we detect eight new candidate bubble collision
signatures for follow-up analysis.
View original:
http://arxiv.org/abs/1202.2861
No comments:
Post a Comment