Normal Parameters for an Analytic Description of the CMB Cosmological Parameter Likelihood

Abstract

The normal parameters are a non--linear transformation of the cosmological parameters whose likelihood function is very well--approximated by a normal distribution. This transformation serves as an extreme form of data compression allowing for practically instantaneous calculation of the likelihood of any given model, as long as the model is in the parameter space originally considered. The compression makes all the information about cosmological parameter constraints from a given set of experiments available in a useable manner. Here we explicitly define the normal parameters that work for the current CMB data, and give their mean and covariance matrix which best fit the likelihood function calculated by the Monte Carlo Markov Chain method. Along with standard parameter estimation results, we propose that future CMB parameter analyses define normal parameters and quote their mean and covariance matrix.