Fitting covariance matrix models to simulations

TL;DR

This paper introduces a likelihood-based method to fit theoretical covariance matrix models to simulation data, reducing the number of simulations needed for reliable covariance estimates in cosmological analyses.

Contribution

It presents a novel approach for fitting covariance models with fewer simulations and provides tests to validate the model's accuracy using $^2$ statistics.

Findings

Model covariance with 2 free parameters matches large-sample covariance accuracy.

Method reduces simulation requirements for reliable covariance estimation.

Partial success in modeling large covariance matrices with improvements over Gaussian assumptions.

Abstract

Data analysis in cosmology requires reliable covariance matrices. Covariance matrices derived from numerical simulations often require a very large number of realizations to be accurate. When a theoretical model for the covariance matrix exists, the parameters of the model can often be fit with many fewer simulations. We write a likelihood-based method for performing such a fit. We demonstrate how a model covariance matrix can be tested by examining the appropriate $\chi^2$ distributions from simulations. We show that if model covariance has amplitude freedom, the expectation value of second moment of $\chi^2$ distribution with a wrong covariance matrix will always be larger than one using the true covariance matrix. By combining these steps together, we provide a way of producing reliable covariances without ever requiring running a large number of simulations. We demonstrate our method on two examples. First, we measure the two-point correlation function of halos from a large set of $10000$ mock halo catalogs. We build a model covariance with $2$ free parameters, which we fit using our procedure. The resulting best-fit model covariance obtained from just $100$ simulation realizations proves to be as reliable as the numerical covariance matrix built from the full $10000$ set. We also test our method on a setup where the covariance matrix is large by measuring the halo bispectrum for thousands of triangles for the same set of mocks. We build a block diagonal model covariance with $2$ free parameters as an improvement over the diagonal Gaussian covariance. Our model covariance passes the $\chi^2$ test only partially in this case, signaling that the model is insufficient even using free parameters, but significantly improves over the Gaussian one.