Entry UTOPIA

[AngusWright]

Angus' Useless Technique which OutPerforms Intelligent Alternatives (UTOPIA)
This is the UTOPIA entry presented at the DESC collaboration meeting. I will comment below with some standard performace.
This entry requires rpy2, and a few R Packages. As part of "the atlantics" we constructed a conda environment that allows rpy2 to run on cori. That is probably best used here as well.

[AngusWright]

Here are two baseline stats run on the Buzzard training sample, using half* the training dataset:

UTOPIA run_5 {'bins': 5, 'sparse_frac': 0.5, 'colors': True, 'errors': False} {'SNR_ww': 247.7985949226061, 'SNR_gg': 1362.251443458753, 'SNR_3x2': 1363.1740218951588, 'FOM_3x2': 60803.64884830707}
UTOPIA run_10 {'bins': 10, 'sparse_frac': 0.5, 'colors': True, 'errors': False} {'SNR_ww': 249.10788992990285, 'SNR_gg': 1674.2085107073099, 'SNR_3x2': 1674.892809715764, 'FOM_3x2': 51118.26862384448}

*Because of this method, a larger training set may perform better still. I will update the metrics when it finishes.