“…However,
is only based on one random split of
, which leads to additional variance due to learning from
instead of
. As proposed by Sanguiao‐Sande and Zhang (
2021), we can reduce the variance of
by applying the Monte Carlo Rao–Blackwell (RB) method. Then, our proposed estimator is given by
where
is the estimator () calculated from the k th Monte Carlo subsamples,
such that
…”