Data compression and covariance matrix inspection: Cosmic shear

Covariance matrices are among the most difficult pieces of end-to-end cosmological analyses. In principle, for two-point functions, each component involves a four-point function, and resulting covariance often has hundreds thousands elements. We investigate various compression mechanisms capable vastly reducing size matrix in context cosmic shear statistics. This helps identify which its parts crucial to parameter estimation. start with simple methods, by isolating "removing" 200 modes associated lowest eigenvalues, then those signal-to-noise ratio, before moving on more sophisticated schemes like at tomographic level and, finally, Massively Optimized Parameter Estimation Data (MOPED). find that, while these approaches prove useful few parameters interest, $\Omega_m$, simplest yield loss constraining power intrinsic alignment (IA) as well $S_8$. For case considered -- from first year data Dark Energy Survey only MOPED was able replicate original constraints 16-parameter space. Finally, we apply tolerance test elements compressed obtained confirm that IA $A_{\mathrm{IA}}$ is susceptible inaccuracies matrix.