Аннотация:Point and interval probability estimates
for an event that has never been observed in a Bernoulli
trial series (0-event) are proposed and validated. In
this case, the classical statistical methods yield a zero
point estimate, which is often unacceptable in practice.
Nonzero point and interval probability estimates for a
0-event are proposed and validated. A classi cation of
samples by size for the case of a 0-event is proposed.
