Running short period planet simulations in supervised case as part of the Quarter 1 data, to test recovery only use half the sample of simulations for supervision and examine the recovery rate of the other half. Would like to see a recovery plot on whatever metric is 'detection' similar to Figure 10 in Schwamb et al 2013 for the sample not part of the training /supervised sample.
