Adding the AETC and AETC-BLUE algorithms in etc.py for mean estimation.
The Jacobian in groupacv.py needed an additional dimension to have the correct shape. This aligns with the current implementation of the Hessian.
I added a flag to the MultiOutputVariance stat which indicates to return the covariance off-diagonal elements of the covariance matrix when estimating the variance. This is necessary for GroupACV to estimate the variances. Let me know if there is another implementation that you prefer.
The tests in test_groupacv.py are currently passing. The only function that does not assert anything is test_groupacv_variance_estimation which is testing the optimization of variance estimation. We can add an assert statement to this soon.
groupacv.pyneeded an additional dimension to have the correct shape. This aligns with the current implementation of the Hessian.test_groupacv.pyare currently passing. The only function that does not assert anything istest_groupacv_variance_estimationwhich is testing the optimization of variance estimation. We can add an assert statement to this soon.