An Analytical Approach to Bayesian Evidence Computation

Bayesian evidence is a key tool in model selection, allowing comparison of models with different numbers parameters. Its use the analysis cosmological has been limited by difficulties calculating it, current numerical algorithms requiring supercomputers. In this paper we give exact formulae for case Gaussian likelihoods arbitrary correlations and top-hat priors, approximate likelihood distributions leading non-Gaussianities (skewness kurtosis). We apply these to without isocurvature components, compare results previously obtained using thermodynamic integration. find that are lower precision than integration, while still being good enough be useful.