leave_one_out

ell1fit.spin_periodicity.leave_one_out(t, ys, variances, min_period, max_period, oversample=5, n_shuffle=1000, rng=None)

Repeat a (single or joint) search with each epoch left out in turn.

A candidate period carried by one or two epochs moves, or loses its significance, when they are dropped; one supported by the whole dataset does neither.

Parameters:

t, ys, variances

As search_joint_periodicity(); a single quantity is a one-element list.

Returns:

list of dict

One per epoch: dropped_mjd, best_period_days, best_power, fap.