1: \begin{abstract}

2: We apply a method recently introduced to the statistical literature to directly

3: estimate the precision matrix from an ensemble of samples drawn from a 

4: corresponding Gaussian distribution. Motivated by the observation that cosmological 

5: precision matrices are often approximately sparse, the method allows one to exploit this

6: sparsity of the precision matrix to more quickly converge to an asymptotic

7: $1/\sqrt{N_{\rm sim}}$ rate while simultaneously providing an error model for all of the

8: terms.

9: Such an estimate can be used as the starting point for further regularization

10: efforts which can improve upon the $1/\sqrt{N_{\rm sim}}$ limit above, and incorporating

11: such additional steps is straightforward within this framework.

12: We demonstrate the technique with toy models and with an example motivated by

13: large-scale structure two-point analysis, showing significant improvements in

14: the rate of convergence.

15: For the large-scale structure example we find errors on the precision matrix

16: which are factors of 5 smaller than for the sample precision matrix for

17: thousands of simulations or, alternatively, convergence to the same error level

18: with more than an order of magnitude fewer simulations.

19: \end{abstract}

20: