Epsilon-sufficient set
= Epsilon-sufficient set
{title2=$\Pr(X\in S)\geq1-\epsilon$}
For a finite-valued <random variable>, an epsilon-sufficient set $S$ retains probability at least $1-\epsilon$. This notion is about covering probability mass and is unrelated to a <sufficient statistic>. For an <IID random variable> source, any such set of words with fixed $\epsilon<1$ requires exponentially many words at a rate approaching the <Shannon entropy>.